2018 Solid-State Lighting R&D Opportunities · Solid-state lighting (SSL), particularly light...

2018 Solid-State Lighting

R&D Opportunities

January 2019

ii

[This page has intentionally been left blank.]

iii

Disclaimer This report was prepared as an account of work sponsored by an agency of the United States Government.

Neither the United States Government, nor any agency thereof, nor any of their employees, nor any of their

contractors, subcontractors, or their employees, makes any warranty, express or implied, or assumes any legal

liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product,

or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to

any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise,

does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States

Government or any agency, contractor, or subcontractor thereof. The views and opinions of authors expressed

herein do not necessarily state or reflect those of the United States Government or any agency thereof.

This publication may be reproduced in whole or in part for educational or non-profit purposes without special

permission from the copyright holder, provided acknowledgement of the source is made. The document should

be referenced as:

DOE BTO SSL Program, “2018 Solid-State Lighting R&D Opportunities,” edited by James Brodrick, Ph.D.

Editor: James Brodrick, DOE BTO SSL R&D Program

Lead Author: Morgan Pattison, SSLS, Inc.

Contributors: Norman Bardsley, Bardsley Consulting

Clay Elliot, Navigant Consulting, Inc.

Monica Hansen, LED Lighting Advisors

Kyung Lee, Navigant Consulting, Inc.

Lisa Pattison, SSLS, Inc.

Jeffrey Tsao, Sandia National Laboratories

Mary Yamada, Navigant Consulting, Inc.

iv

[This page has intentionally been left blank.]

v

List of Acronyms Abbreviation Definition

$/klm U.S. dollars per kilolumen

3-D 3-dimensional

A/cm2 amperes per square centimeter

AC alternating current

ALD atomic-layer deposition

AlGaN aluminum gallium nitride

AlInGaN aluminum indium gallium nitride

AlInGaP aluminum indium gallium phosphide

AlN aluminum nitride

ANSI American National Standards Institute

app software application

ASP average selling price

BLE Bluetooth Low Energy

BR bulged reflector

BTO Building Technologies Office

Btu British thermal unit

°C degrees Celsius

CALiPER Commercially Available LED Product Evaluation and Reporting

CCT correlated color temperature

cd/m2 candelas per square meter

CFL compact fluorescent lamp

CIE Commission Internationale de l’Eclairage

CLTB connected lighting test bed

cm-LED color-mixedLED

COB chip-on-board

CRI color rendering index

CSA China Solid State Lighting Alliance

CSP chip scale package

CWF cool white fluorescent

DC direct current

DLC DesignLights Consortium

DMX digital multiplex

DOE U.S. Department of Energy

Duv distance from the blackbody locus in u-v colorspace

EESL Energy Efficiency Services Limited

EQE external quantum efficiency

EQE/IQE extraction efficiency/internal quantum efficiency

EU European Union

FCC Federal Communications Commission

FOA funding opportunity announcement

vi

FP-7 Seventh Framework Programme

FWHM full width at half maximum

FY fiscal year

GaAs gallium arsenide

GaN gallium nitride

HID high intensity discharge

HPS high-pressure sodium

HVAC heating, ventilation and air conditioning

hy-LED hybrid LED

IEA International Energy Agency

IES Illuminating Engineering Society

IHS Information Handling Services Markit Ltd.

III-N III nitride material

III-V III-V semiconductor material

InGaN indium gallium nitride

IoT Internet of things

IP Internet protocol

IQE internal quantum efficiency

IR infrared

ITO indium tin oxide

K Kelvin

klm/m2 kilolumen per square meter

KrW Korean Won

L1 Level 1

L2 Level 2

L70 duration of lumen maintenance to 70% initial brightness; operational lifetime

LBNL Lawrence Berkeley National Laboratory

LCA life-cycle assessment

LED light-emitting diode

LER luminous efficacy of radiation

lm/m2 lumens per square meter

lm/W lumens per watt

LSRC LED Systems Reliability Consortium

LT50 lifetime to 50% of the initial luminance

LT80 lifetime to 80% of the initial luminance

mAh milliamp hour

MC-PCB metal-core printed circuit board

MEMS microelectromechanical systems

Mg magnesium

MLA micro-lens array

MOCVD metal organic chemical vapor deposition

MR multifaceted reflector

MW megawatt

vii

NGLIA Next Generation Lighting Industry Alliance

NIST National Institute of Standards and Technology

nm nanometer

OLED organic light emitting diode

OVPD organic vapor phase deposition

PAR parabolic aluminized reflector

PCB printed circuit board

PCE power conversion efficiency

PCS peelable-clean surface

pc-LED phosphor-converted LED

PDMS polydimethylsiloxane

Pd palladium

PECVD plasma enhanced chemical vapor deposition

PLC powerline communication

PNNL Pacific Northwest National Laboratory

PoE power over Ethernet

PPER photosynthetic photon efficacy of radiation

Pt platinum

QY quantum yield

R&D research and development

R2R roll-to-roll

R9 Color fidelity test standard for red content not used in calculations of CRI

RGB red, green and blue

RGBA red, green, blue and amber

RYGB red, yellow, green and blue

R&D research and development

SEMLA sub-electrode microlens array approach

SiC silicone carbide

SPP surface plasmon polariton

SSL solid-state lighting

TADF thermally activated delayed fluorescence

TAKT process cycle time

TBtu trillion British thermal units

THD total harmonic distortion

TiO2 titanium dioxide

TLED tubular LED

TWh terawatt-hours

UDC Universal Display Corporation

UV ultraviolet

V volt

VLC visible light communication

W Watt

W/m2 watts per square meter

viii

W/mK watts per meter kelvin

W/mm2 watts per square millimeter

YAG yttrium aluminum garnet

ZESCO Zambia Electricity Supply Corporation

ZrO2 zirconium dioxide

µm micrometer

Δu'v' magnitude of color shift in the CIE 1976 chromaticity diagram (u', v')

Ω/ resistivity per unit area

ix

Executive Summary Solid-state lighting (SSL), particularly light emitting diode (LED) based SSL, is on course to become the

dominant technology across all lighting applications. The luminous efficacy, as measured in lumens per Watt

(lm/W) of SSL continues to advance toward the practical limit of 255 lm/W for phosphor converted LED

architectures and the ultimate theoretical limit of 325 lm/W for direct emitting architectures. These

advancements are the result of numerous and ongoing breakthroughs in fundamental, early stage R&D that

have been applied across the SSL value chain. This document provides further detail on these advancements

and priority R&D topics suggested by members of the U.S. lighting science R&D community who collaborate

with the U.S. Department of Energy’s (DOE) Building Technologies Office (BTO), within the Office of

Energy Efficiency and Renewable Energy, on its SSL Program.

High efficacy and low costs result in efficient and cost-effective lighting that is rapidly being adopted and

saving substantial amounts of energy. In addition, unlike previous energy saving lighting technologies, LED

lighting does not require compromises in performance. LED lighting can be engineered to have almost any

spectrum that can provide good color quality. It is fundamentally dimmable, both instantaneously and

continuously. Finally, the small source size enables improved optical control. These performance features,

coupled with the high efficacy and low cost of LED lighting, enable a rare trifecta for energy efficiency

technologies – high efficiency, low cost, and improved performance – which has resulted in rapid adoption and

massive energy savings. These features and benefits can also be achieved while reducing the environmental

impacts of lighting (beyond benefits from reduced energy generation) in terms of reduced materials toxicity

and ecological impacts.

So far, the benefits resulting from the transition to SSL have accrued through replacement of conventional

lighting products to more efficient SSL products for the same basic lighting job – illumination and visibility

based on the photopic eye response. However, the new capabilities of SSL technology, coupled with new

understanding in lighting science, open up possibilities to further reduce lighting energy consumption, improve

lighting performance in new ways, and reduce negative impacts of earlier lighting technologies.

x

Beyond improvements in efficiency or efficacy offered by SSL technology, further energy savings can be

achieved through improved optical control, spectral tailoring, and more precise control of intensity. However,

a new framework for modeling and evaluating trade-offs between these factors, as well as source efficiency,

needs to be developed. Spectral engineering enables improved lighting performance by offering the ability to

tailor light for very specific features such as color gamut, discrimination, and replication. Improved optical

control could enable reduced glare and more precise delivery of light. Controlling the intensity also enables

lighting to achieve the desired levels by application, and to adapt dynamically in response to changes in the

outside environment or human needs.

Recent research and an improved understanding of lighting science has shown that there are considerations

beyond basic illuminance and color qualities (such as color rendering index (CRI) or correlated color

temperature (CCT)). Light that is intense and has a higher blue content affects alertness and melatonin

secretion in humans leading to significant health implications. Lighting designers will need to consider these

effects as well as photopic eye response when specifying new lighting installations. Lighting product

developers will need clear guidance in terms of intensity and spectrum to develop lighting products that are

both efficient and supportive of health or well-being compared to legacy lighting installations. This rare

combination of energy saving and well-being benefits is offered solely by SSL technology.

The bundle of new discoveries, impacts, and potential benefits of SSL stem largely from advancements in LED

efficiency. Improvements to efficiency reduce costs and size while enabling new form factors, improved

optical performance, and new lighting applications. Ongoing research and development (R&D) in SSL

materials and device efficiency is therefore necessary to meet projected energy savings targets and to also

maximize downstream energy savings through improved lighting application efficiency. Future improvements

in efficiency for new lighting applications must ensure well-being and productivity are not compromised, and

even improved where possible.

Finally, given the fast pace of SSL development and improvement, it is critical that manufacturing approaches

keep pace by improving throughput, increasing flexibility, reducing materials usage, and generally advancing

manufacturing speed and efficiency. In addition, within the U.S., the vast range of lighting product types and

application requirements means that increasingly, products will be customized and produced ‘on-demand’. To

support this trend, new flexible manufacturing approaches are necessary. In particular, developing additive

manufacturing techniques to maximize the range of products while simultaneously minimizing stored

component inventory could have a large impact. Additive manufacturing techniques could be deployed across

the manufacturing value chain, from wafer level processing techniques to light fixture creation.

Another technology group within the SSL family is organic LED (OLED). OLED products have the potential

to offer unique benefits complementary to LED lighting. However, significant technology barriers remain for

OLED lighting, with progress lagging behind LED performance and cost. OLED efficacy greatly lags LED

efficacy at approximately 90 lm/W with a target of 190 lm/W; however, there may be application-specific

advantages. OLED lighting technology needs ongoing R&D to translate lab scale efficiency and performance

advancements to commercially practical approaches. In particular, efficiency can be improved at the material,

device, and light extraction levels. Additionally, these advancements in efficiency would have a direct impact

on cost and reliability.

OLED lighting offers an intriguing performance and production counter-point to LED lighting. By its very

nature, OLED lighting is diffuse, meaning it can be placed very close to the occupant or object being lit. Every

other lighting technology, including LEDs, requires optical diffusion to protect occupants from glare from the

bright light source. Now, many LED lighting products use waveguide optics to achieve a flat profile, while

OLED products such as those shown in Figure 1-2, inherently provide diffuse lighting without the need for

additional optics. While OLED lighting offers these unique lighting prospects, OLED lighting does not yet

readily work with pre-existing, ubiquitous lighting form factors.

xi

Figure 1-1. OLED Panel

Source: Digital Trends, September 2018 [1]

In terms of production, OLEDs require nanometer scale control of organic material deposition thickness over

very large areas (square meters) at high speed. This level of production control is challenging and requires the

development of new manufacturing technologies that are compatible with the most efficient OLED materials

and device approaches. In addition, OLED devices and materials must be protected from environmental

incursions of oxygen and moisture that can disrupt the finely tuned material and device performance.

Furthermore, these technology and manufacturing challenges must be achieved while meeting consumer

demands for cost, performance, and reliability. If these barriers are overcome, significant benefits to lighting

applications could be achieved in terms of energy savings and human comfort and overall well-being. Any

advances in OLED lighting would also aid the development of a broad range of similar applications including

photovoltaics, displays, advanced fenestration, and beyond – with the potential for significant additional cross-

cutting energy savings.

This document discusses R&D topics necessary to make advancements in the areas described above. The

specific research topics listed are the result of BTO experts in consultation with stakeholders from academia,

national laboratories, and industry who work with BTO’s SSL Program. DOE will make the final

determination of R&D topics for possible funding. BTO’s SSL Program continues to set aggressive

performance and cost targets for LED and OLED lighting technology and provide early stage R&D funding

and research activities in support of reaching these targets.

The current critical R&D challenges identified by stakeholders during the Roundtable and Workshop

discussions are listed below. Further description of the specific R&D topics with supporting metrics is given in

Section 3.

xii

LED-Based Lighting R&D

• Light Emitting Diode Devices and Materials – Push LED emitters across the visible spectrum that

demonstrate advancements in peak efficiency and stable efficiency at high current drive and temperature

operating conditions. Also, development of fundamental models to predict LED device performance

across a range of materials, device structures, and synthesis techniques.

• Advanced Emitter Device Architectures – Explore the use of advanced emitter device architectures

with state-of-art-emitter materials to improve the extraction of white photons from a device package, as

measured by overall package power conversion efficiency (lm/W), and the ability to deliver white

photons to a target, which generally improves with luminous emittance (lm/mm2).

• Quantum Dot Optical Down-Converters – Research to advance understanding in high efficiency, on-

chip quantum dot (QD) down converters to match or exceed performance of conventional on-chip

phosphor materials at a range of emission wavelengths.

• Advanced LED Lighting Concepts – Develop fully optimized color-mixed direct LED lighting product

concepts that demonstrate efficiency advancements or lighting products with advancements in lighting

application efficiency.

• LED Power and Functional Electronics – Develop advanced prototype power delivery concepts for

luminaires with high efficiency across the operating range, high reliability, and minimal size and weight.

• Additive Fabrication Technologies for Lighting – Develop high volume additive manufacturing

technologies for any portion of the LED lighting manufacturing value chain that reduce part count and

are cost effective.

OLED-Based Lighting R&D

• Stable, Efficient White Devices – Develop novel materials and structures that can help create a highly

efficient, stable white OLED device.

• Advanced Fabrication Technology for OLEDs – Develop novel approaches and advancements in

materials deposition, device fabrication, or encapsulation of OLED panels that lead to significant

reduction in processing costs without degrading performance.

• Light Extraction and Utilization – Devise new optical and device designs for improving OLED light

extraction while retaining the thin profile and advancing state-of-the-art performance of OLED panels.

• OLED Prototype Lighting Platforms – Develop OLED lighting platforms that achieve the

performance and design goals of OLED lighting technology and demonstrate clear differentiation from

existing products.

Cross-Cutting Lighting R&D (LED and OLED)

• Understanding Human Physiological Impacts of Light to Improve Efficiency– Research to understand

and define physiologically optimized lighting for the general population based on objective physiological

responses to light and/or large-scale collection and review of subjective responses for the purpose of

optimizing lighting efficiency.

• Understanding Lighting Application Efficiency – Develop a general framework, mathematical model,

and computer simulation approach to characterize lighting application efficiency for any lighting

application in terms of the four primary aspects of lighting application efficiency: light source efficiency,

optical delivery efficiency, spectral efficiency, and intensity efficacy.

xiii

Table of Contents Executive Summary ................................................................................................................................... ix

Introduction .................................................................................................................................................. 1

Research Topics .......................................................................................................................................... 2

3.1 Process and Discussion ................................................................................................................... 2

3.1.1 Goals and Projections ........................................................................................................... 2

3.2 LED Research Needs ...................................................................................................................... 6

3.2.1 LED Research Tasks ............................................................................................................ 7

3.3 OLED Research Needs ................................................................................................................. 12

3.3.1 OLED Research Tasks ....................................................................................................... 12

3.4 Cross-Cutting Lighting R&D (LED and OLED) .......................................................................... 15

Directions in Lighting Science ................................................................................................................ 17

4.1 Lighting Application Efficiency – A New Framework and Opportunity for Energy Savings ..... 19

4.1.1 Spectral Efficiency ............................................................................................................. 19

4.1.2 Source Efficiency ............................................................................................................... 20

4.1.3 Optical Delivery Efficiency ............................................................................................... 21

4.1.4 Intensity Effectiveness ....................................................................................................... 22

4.2 Visual and Non-Visual Responses to Light .................................................................................. 22

4.2.1 R&D Directions for Visual and Non-Visual Responses to Light....................................... 23

4.2.2 Understanding Relationship Between Energy Savings and Wellness Implications ........... 24

4.3 Connected Lighting ...................................................................................................................... 24

4.3.1 Energy Savings and Other Valued Features ....................................................................... 25

4.3.2 Lighting Controls Interoperability ..................................................................................... 26

4.3.3 Connected Lighting Test Bed ............................................................................................. 26

4.3.4 Security .............................................................................................................................. 27

Directions in LED Science ....................................................................................................................... 28

5.1 White LED Technology ................................................................................................................ 28

5.2 Advanced Material Discovery and Engineering Exploration for LED Development .................. 31

5.3 Understanding Droop in LEDs ..................................................................................................... 31

5.3.1 Blue LEDs .......................................................................................................................... 32

5.3.2 Green LEDs........................................................................................................................ 33

5.3.3 Thermal Droop ................................................................................................................... 34

5.4 Advanced LED Architectures ....................................................................................................... 35

5.4.1 Droop Mitigation ................................................................................................................ 35

5.4.2 High Luminance ................................................................................................................. 37

5.5 Optical Down-Converters ............................................................................................................. 38

5.5.1 Narrow-Band Phosphors .................................................................................................... 39

5.5.2 Quantum Dot Down-Converters ........................................................................................ 40

5.6 Additive Fabrication Technologies for Lighting .......................................................................... 42

5.7 Advanced LED Lighting Concepts ............................................................................................... 44

5.8 LED Power and Functional Electronics ....................................................................................... 45

5.8.1 Driver Performance ............................................................................................................ 45

xiv

5.8.2 Reliability ........................................................................................................................... 46

5.8.3 Enhanced Functionality of Drivers .................................................................................... 47

Directions in OLED Science .................................................................................................................... 48

6.1 Stable, Efficient White Organic Emitters ..................................................................................... 49

6.2 Light Extraction ............................................................................................................................ 52

6.2.1 Scattering Layers ................................................................................................................ 53

6.2.2 Functionalized Substrates................................................................................................... 53

6.2.3 Corrugated Substrates ........................................................................................................ 54

6.2.4 Refractive Index ................................................................................................................. 54

6.2.5 Orientation of Emitter Dipoles ........................................................................................... 54

6.2.6 Stack Optimization ............................................................................................................. 55

6.2.7 Light Extraction Enhancement in Flexible OLEDs ........................................................... 55

6.3 Advanced Fabrication Technology for OLEDs ............................................................................ 56

6.3.1 Depreciation Costs ............................................................................................................. 56

6.3.2 Deposition of Organics ...................................................................................................... 57

6.3.3 Substrates and Encapsulation ............................................................................................. 57

6.3.4 Transparent Conductors ..................................................................................................... 58

6.3.5 Substrate Handling ............................................................................................................. 58

6.4 OLED Lighting Platforms ............................................................................................................ 59

6.4.1 OLED Features .................................................................................................................. 59

6.4.2 Special Applications .......................................................................................................... 59

6.4.3 Modules and Light Engines ............................................................................................... 60

6.4.4 Drivers and Power Supplies ............................................................................................... 60

Appendices ................................................................................................................................................ 61

7.1 LED Supply Chain........................................................................................................................ 61

7.1.1 LED Package Manufacturing ............................................................................................. 62

7.1.2 LED Luminaire Manufacturing .......................................................................................... 64

7.1.3 OLED Manufacturing ........................................................................................................ 65

7.2 BTO Program Status ..................................................................................................................... 66

7.2.1 Current SSL Portfolio ........................................................................................................ 66

7.2.2 Patents ................................................................................................................................ 69

Bibliography .............................................................................................................................................. 70

xv

List of Figures Figure 1-1. OLED Panel.................................................................................................................................................................... xi

Figure 3-1. Efficacies and Efficiencies Over Time of White and Colored LED Packages.............................................................. 3

Figure 3-2. White-Light OLED Panel Efficacy Projections............................................................................................................... 5

Figure 4-1. Human Eye Sensitivity Function CIE 1978 Photopic Vision ......................................................................................18

Figure 4-2. Possible Framework for Evaluating Lighting Application Efficiency .........................................................................19

Figure 4-3. Action Spectra for Humans and Plants: (a) Human Action Spectra [4] (b) Plant Action Spectra [5] ......................20

Figure 5-1. Schematic of Two Main White LED Architectures .....................................................................................................28

Figure 5-2. Typical Simulated Spectral Power Density for White-Light LED Package Architectures..........................................29

Figure 5-3. Efficacies and Efficiencies Over Time of White and Colored LED Packages............................................................30

Figure 5-4. Blue LED EQE vs. Current Density and Schematic of LED Quantum Well Valence Band [14] ................................33

Figure 5-5. Spectral Power Densities of State-of-the-Art Commercial LEDs vs. Wavelength .....................................................33

Figure 5-6. Schematic of the LED Quantum Well Valence Band [14] .........................................................................................34

Figure 5-7. LED Efficiency as Function of Junction Temperature [16] ........................................................................................35

Figure 5-8. PCE Vs. Current Density for a State-of-The-Art LED and LD Emitting at Violet Wavelengths [18] ..........................36

Figure 5-9. Schematic Band Diagram of a Stacked Active Region LED with Tunnel Junctions (left); the External Quantum

Efficiency at the Peak LED Operating Current (right) [19] ...........................................................................................................37

Figure 5-10. Luminous Efficacy vs. Luminous Emittances for State-of-the-Art Commercial White LEDs ..................................38

Figure 5-11. Spectrum Comparison of 90 CRI PC-LED, 90 CRI Conventional Red Phosphor LED, and Human Eye Response

[20] .................................................................................................................................................................................................39

Figure 5-12. Light Loss of Phosphors in LED Packages under High Blue Flux Densities (left) and Color Shift under Stressed

Operating Conditions (right) [21] ..................................................................................................................................................40

Figure 5-13. Emission Wavelength of CdSe QDs as a Function of Dot Diameter [26] ...............................................................41

Figure 5-14. Images of 3D-Printed Lighting Fixtures [27] ...........................................................................................................43

Figure 5-15. Deposition of Droplets by UV Print Head onto Substrate Material (left-top). Droplets of Polymer Are Allowed to

“Flow” under Surface Tension before Curing with UV Light, giving Smooth Surfaces Needed for Optics (left-bottom). Array of

Micro-Optic Lenses (right). [28] .....................................................................................................................................................43

Figure 6-1. Acuity OLED Luminaires in Office Space....................................................................................................................48

Figure 6-2. Cynora’s Illustration of TADF as Compared with Fluorescent and Phosphorescent Approaches [36] ...................50

Figure 6-3. Comparison of the Mechanisms of TADF and Hyperfluorescence [41] ...................................................................51

Figure 6-4. University of Michigan Sub-electrode Microlens Array Approach: (a) Device Architecture; (b) Light Extraction

Enhancement Factor [46] .............................................................................................................................................................53

Figure 6-5. External Quantum Efficiency of Phosphorescent OLEDs with Pt-Based Emitters [51] ............................................54

Figure 6-6. OLEDWorks Brite 3 Curve, BendOLED on Corning Willow Glass [57] ......................................................................56

xvi

List of Tables Table 3-1. Phosphor-Converted and Color Mixed LED Package Historical and Targeted Efficacy .............................................. 4

Table 3-2. OLED Panel Historical and Targeted Efficacy ............................................................................................................... 5

Table 3-3. Assumptions for Wavelength and Color as Used in the Task Descriptions ................................................................. 6

Table 3-4. Emitter Materials ............................................................................................................................................................ 7

Table 3-5. Advanced Emitter Device Architectures ........................................................................................................................ 8

Table 3-6. Understanding Quantum Dot Optical Down-Converters ............................................................................................... 9

Table 3-7. Advanced LED Lighting Concepts ................................................................................................................................10

Table 3-8. LED Power and Functional Electronics .......................................................................................................................11

Table 3-9. Additive Fabrication Technologies for Lighting ...........................................................................................................11

Table 3-10. Stable, Efficient White Devices .................................................................................................................................12

Table 3-11. Advanced Fabrication Technology for OLEDs ...........................................................................................................13

Table 3-12. Light Extraction and Utilization ..................................................................................................................................13

Table 3-13. OLED Prototype Lighting Platforms ...........................................................................................................................14

Table 3-14. Understanding Human Physiological Impacts of Light ............................................................................................15

Table 3-15. Understanding Lighting Application Efficiency .........................................................................................................16

Table 5-1. Phosphor-Converted and Color Mixed LED Package Historical and Targeted Efficacy ............................................30

Table 6-1. OLED Historical and Targeted Luminaire Efficiency ...................................................................................................49

Table 6-2. Current Status and Cost Targets for Panels Produced by Traditional Methods ........................................................56

1

Introduction The administration of President Trump calls for investment in “next-generation energy technologies to

efficiently convert them into useful energy services (e.g., light, heat, mobility, power, etc.)” [2]. The U.S.

Department of Energy (DOE) Building Technologies Office’s (BTO) Solid-State Lighting (SSL) Program is

playing a critical role in advancing this agenda by

investing in SSL R&D. The Administration also calls

for “early-stage, innovative technologies that show

promise in harnessing American energy resources

safely and efficiently” [2]. SSL technology is a prime

example of how energy can be converted into a useful

service, namely lighting. SSL has demonstrated clear

advancements in efficiency over incumbent lighting

technologies, resulting in significant energy savings.

SSL still has room to become even more efficient and

has the potential to improve safety by providing light

that reduces the physiological impacts we currently

experience with legacy sources. SSL has the potential

to harness American energy resources to provide light

safely and efficiently.

The BTO SSL Program was created in response to

Congressional direction described in Section 912 of the

Energy Policy Act of 2005, which directs DOE to

“Support research, development, demonstration, and commercial application activities related to advanced

solid-state lighting technologies based on white light-emitting diodes.” The BTO SSL Program has

developed a comprehensive R&D strategy to support

advancements in SSL technology and maximize energy

savings. The specific goal of the R&D Program is:

By 2030, develop advanced SSL technologies that –

compared to conventional lighting technologies – are

much more energy efficient, longer lasting, and cost

competitive, by targeting a product system efficiency

of 50% with appropriate application spectrum.

In order to maximize energy savings, the BTO SSL Program supports foundational R&D topics with benefits

that apply across the value chain and such R&D that is not typically undertaken within the lighting industry.

BTO-supported R&D advances the understanding of underlying physical phenomena, explores new technical

and fabrication approaches, reduces the development risk with new technologies, and/or develops

understanding of application requirements that improve lighting effectiveness, while simultaneously improving

efficiency.

This document, BTO’s 2018 SSL R&D Opportunities, is updated annually and provides analysis, context, and

direction for ongoing R&D activities to advance SSL technology and increase energy savings. Research areas

in this document come from DOE BTO experts with input from members of the lighting science R&D

community at National Laboratories and academia as well as large and small businesses. The inputs are

collected at the BTO SSL Roundtable meetings, the OLED stakeholder meetings, and the annual BTO SSL

R&D Workshop.

American Energy Dominance

“Fueling America’s greatness requires access

to domestic sources of clean, affordable, and

reliable energy. Unleashing these abundant

energy resources will require investment in

next-generation energy technologies to

efficiently convert them into useful energy services (e.g., light, heat, mobility, power,

etc.). Agencies should invest in early-stage,

innovative technologies that show promise in

harnessing American energy resources safely

and efficiently. Federally funded energy R&D

should continue to reflect an increased reliance

on the private sector to fund later-stage

research, development, and commercialization

of energy technologies. Agencies should invest

in user facilities that can improve collaboration

with industry and academia and achieve

advancements across the full spectrum of

discovery, from incremental improvements to

game changing breakthroughs.”

FY 2020 Administration Research and

Development Budget Priorities, July 31, 2018

Mick Mulvaney, OMB, Michael Kratsios,

OSTP

2

Research Topics To reach the full potential of solid-state lighting (SSL), further research is necessary. Despite rapid progress,

there are still significant advancements in performance and scientific understanding that can be made. In terms

of U.S. energy savings from SSL, 90% of the potential remains untapped.1 Advancements in SSL technology

have highlighted gaps in understanding at not only the material-device level, but also at the lighting science

level. Research in these areas will enable the next level of performance advancements for SSL. At the

materials and device level, ongoing innovation and breakthroughs in materials, devices, advanced fabrication

processes, and integration are needed to realize the full potential of the technology. In addition, at the lighting

science level, the SSL technology platform raises new questions as to the effectiveness of the delivery and

control of lighting, as well as the effectiveness of lighting for engaging both visual responses and non-visual

physiological responses.

The suggested R&D topics described in this document are inputs given by members of the U.S. lighting

science R&D community who collaborate with the BTO SSL R&D Program. These stakeholders include

academic, National Laboratory, and industry researchers who provide feedback and inputs to the BTO SSL

Program. These topics do not represent forward looking directions by the BTO SSL Program, but rather

stakeholder suggestions as to the most critical areas for advancement of SSL technologies now. Further

analysis and discussion of the suggested R&D topics is provided in the subsequent sections of this document.

3.1 Process and Discussion

The BTO SSL Program has responded to the SSL opportunity by providing direction and coordination of

multiple R&D efforts intended to advance the technology and to promote the ultimate energy savings offered

by the technology.2 The BTO SSL Program seeks to fund research that offers innovative advancements such as

unique materials or device structures, next generation integration concepts, or develops a new understanding of

the underlying technology and lighting science to provide safe and efficient lighting.

3.1.1 Goals and Projections

This section describes expectations for progress toward BTO SSL efficiency goals over time based on

performance to date. The projections are based on best-in-class performance, normalized to particular

operating conditions to track progress.3 These advancements translate to improved performance industry wide

and promote domestic leadership in this technology.

Efficiency and Efficacy Projections for LEDs

Figure 3-1 panel (a) uses a logistic fit to project efficacy over time for cool white and warm white phosphor

converted LED packages and color mixed LED packages. An upper limit of 250 lm/W is assumed for

phosphor converted LED packages and an upper limit of 325 lm/W is assumed for color mixed packages.4

Panel (b) of Figure 3-1 shows projections for power conversion efficiency of blue (440-460 nm), green (530-

550 nm), amber (570-590 nm), and near red (610-620 nm) direct emitting LEDs, again with a logistic fit for

projected performance, and with an upper limit of 90% power conversion efficiency. Table 3-1 shows

historical and projected LED package efficacy for warm white and cool white phosphor converted-LEDs (PC-

LEDs), and color mixed LEDs (cm-LEDs). The assumed operating conditions for qualified data points may

not correspond to practice, particularly with respect to the increasing use of lower drive currents to minimize

current droop. Nevertheless, using a standard current (or power density) at a fixed operating temperature and

1 DOE Solid-State Lighting Program, “Adoption of Light-Emitting Diodes in Common Lighting Applications,” Prepared by Navigant Consulting Inc.,

July 2017. https://www.energy.gov/sites/prod/files/2017/08/f35/led-adoption-jul2017_0.pdf 2 For more information on the DOE SSL Program see: https://energy.gov/eere/ssl/about-solid-state-lighting-program 3 In practice, by adjusting the operating conditions at the luminaire level, LED performance can be improved or degraded from the normalized

operating conditions used for these projections. 4 Updated from analysis in DOE Solid-State Lighting Program, “2016 R&D Plan,” Prepared by Navigant Consulting Inc., June 2016.

https://www.energy.gov/sites/prod/files/2018/09/f56/ssl_rd-plan_jun2016.pdf.

https://www.energy.gov/sites/prod/files/2017/08/f35/led-adoption-jul2017_0.pdf

https://energy.gov/eere/ssl/about-solid-state-lighting-program

https://www.energy.gov/sites/prod/files/2018/09/f56/ssl_rd-plan_jun2016.pdf

3

selecting devices within limited ranges of CCT and CRI allow researchers to evaluate developments in emitter

efficiency (including the reduction of current and thermal droop) and down-converter performance.

Figure 3-1. Efficacies and Efficiencies Over Time of White and Colored LED Packages

All curves are logistic fits using various assumptions for long-term future performance and historical experimental data.

The data are from qualified products at the representative operating conditions of 25°C and 35 A/cm2 input current

density. They will differ from some commercial products, particularly those that operate at lower drive current densities to

minimize current droop.

The upper panel (a) are the luminous efficacies of warm white (3000 K) and cool white (5700 K) phosphor-converted

LEDs and hypothetical color-mixed LEDs (CM-LEDs) with a CCT of 3000-4000 K. Luminous efficacies have the typical units

of photopic lumens of light (lm) created per input electrical Watt (We) of wall-plug power. Year 2017 commercial products

reach approximately 180 lm/W for cool white PC-LEDs and approximately 160 lm/W for warm white PC-LEDs. These

values correspond to raw electrical-to-optical power-conversion efficiencies of approximately 0.5 Wo/We.

The lower panel (b) are the power-conversion efficiencies of direct-emitting LEDs at the various colors (blue, green, amber,

and near-red) necessary for CM-LED white light of highest source luminous efficacy and high color rendering quality.

Approximate future potential power-conversion efficiencies are depicted as a saturation at 90% for all colors beginning in

the years 2035–2040. The historical power conversion efficiencies of these sources were combined and appropriately

weighted to give the CM-LED LEDs and conversion efficiencies depicted in the upper panel (a).

4

Table 3-1. Phosphor-Converted and Color Mixed LED Package Historical and Targeted Efficacy

Metric Type 2016 2017 2025 2035 Final Goal

LED Package Efficacy

(lm/W)

PC Cool

White 160 167 241 249 250

PC Warm

White 140 153 237 249 250

Color Mixed 90 100 196 288 325

Figure 3-1 and Table 3-1 track pc-LED package progress, as that is the mainstream package architecture used

in SSL products. Alternative approaches, such as a hybrid combination of a red LED and a phosphor-converted

LED, could meet the asymptote more quickly than pc-LEDs due to the availability of narrow linewidth red

LED sources.

Efficacy Projections for OLEDs

OLED efficacy has been improving, but not at the desired pace. Though material and device technology has

been demonstrated in the laboratory for achieving OLED panels of much greater than 100 lm/W, low light

extraction efficiency remains a key technical challenge. Integrating light extraction technology without

disrupting the yield and stability of devices has presented a major challenge. Figure 3-2 and Table 3-2 project

OLED panel efficacy based on past performance and anticipated progress. The dashed curve presents BTO

OLED panel efficacy goals put forth in 2014, whereas the solid curve shows projections based on performance

to date. It was realized during the intervening three years that many of the techniques used to raise the efficacy

of laboratory devices could not be implemented quickly in commercial products manufactured in high volume.

Even though panels with efficacy of 90 lm/W were listed in product catalogs in 2017, they were available only

in small quantities.

The changes in these forecasts reflect strategic decisions made by the industry around 2015 to prioritize color

quality over efficacy. Increased lifetime and reliability were also prioritized over efficacy. Having met goals

with respect to color quality and lifetime, researchers are once again focusing on efficacy with a goal of 190

lm/W. Increased light extraction will be key to success in this endeavor. While significantly lower than the

LED efficacy goal of 325 lm/W, OLEDs may be able to offer application specific advantages requiring less

light production overall, and therefore lower energy use, for specific use cases.

Data on OLED panels remain rather sparse and show a lot of variation, so there is considerable uncertainty in

the projected curve. There is also a significant difference in the efficacy of rigid and flexible panels. Qualified

points were defined as those for panels with a minimum area of 50 cm2, CRI greater than or equal to 80, and

CCT between 2580 K and 3710 K. The average of qualified data for each year was used to fit the projection

curve.

5

Figure 3-2. White-Light OLED Panel Efficacy Projections

Table 3.2 summarizes a path toward achieving an efficacy of 190 lm/W with low rates of lumen depreciation.

Table 3-2. OLED Panel Historical and Targeted Efficacy

Metric 2016 2017 2025 2035 Goal

Panel efficacy

(lm/W)* 60 90 160 180 190

* Projections assume CRI > 80, CCT = 2580K- 3710K.

Achieving efficiency gains and lumen depreciation goals will not alone be sufficient to make meaningful

advancements for OLED lighting. The films must also be consistently manufacturable in large areas at low

cost, which may limit material choices and stack configurations. Improvements to the stability of OLED

luminaires must also be realized. OLEDs are sensitive to oxygen, moisture, and other pollutants in the

operating environment. Thus, extensive encapsulation of the OLED panel is required, particularly on flexible

substrates. In addition, oxygen, moisture, and other contaminants can become embedded into the OLED in the

fabrication process, reducing the panel lifetime.

With respect to reliability, the current specifications on lumen depreciation, L70 at 40,000 to 100,000 hours at

3000 cd/m2, seem sufficient for almost all applications. The emphasis needs to change to reducing the

probability of catastrophic failure, unacceptable color shift, or voltage increases within the anticipated lifetime

of the light.

6

3.2 LED Research Needs

Specific task tables in subsequent LED research sections reference color or descriptive terms for color

temperature. Table 3-3 shows these ranges for various color wavelengths and explains the meaning of color

temperature.

The milestones provided in the tasks described below represent the minimal descriptions for progress. They

provide initial and interim targets for quantitative evaluation of progress. All these tasks will require some

additional system-level performance description and, most likely, additional metrics specific to the proposed

approach. Researchers in these areas are expected to possess and communicate a detailed, system-level

understanding of the role of the described research. Where appropriate, researchers should further define and

describe metrics and milestones that are necessary to demonstrate progress in the research topic.

Table 3-3. Assumptions for Wavelength and Color as Used in the Task Descriptions

Color Dominant Wavelength or CCT CRI

Blue 440-460 nm N/A

Green 530-550 nm N/A

Amber 570-590 nm N/A

Near Red 610-620 nm N/A

Warm White 3000 K ≥ 80

Cool White 5700 K ≥ 70

7

3.2.1 LED Research Tasks

Table 3-4. Emitter Materials

Light Emitting Diode Devices and Materials

Description: Develop new or improved emitter materials with an advanced fundamental understanding of materials-

synthesis-performance relationships for light emitting diodes. Research includes theoretical analysis, analysis of historical

results, experimental results, and deep characterization in a closely structured experiment designed to yield more definitive

scientific understanding. Project results should enable some of the following:

• Guidance for improving red, amber, and green LED performance.

• General understanding and model for prediction of LED performance in different materials systems.

• Fundamental understanding of current density and thermal droop that can enable improved mitigation

approaches.

• Modeling that can help project device performance from materials properties of new emitter material systems.

Work on novel LED materials should demonstrate a path toward meeting 2025 performance targets. All research should be

on highest caliber materials and devices to yield clearest possible results. Results should be impactful for the application of

energy saving solid-state lighting by defining a path to achievement of ultimate BTO SSL performance targets described in

table below.

Metrics 2017 Status Interim 2025 Targets 2035 Targets

EQE (peak value)

80% (Blue)

44% (Green)

63% (Near Red*)

18% (Amber*)

88% (Blue)

60% (Green)

69% (Near Red)

33% (Amber)

93% (Blue)

75% (Green)

80% (Near Red)

60% (Amber)

PCE† - 35A/cm2, 25oC

67% (Blue)

27% (Green)

50% (Near Red*)

16% (Amber*)

84% (Blue)

50% (Green)

70% (Near Red)

30% (Amber)

90% (Blue)

75% (Green)

85% (Near Red)

70% (Amber)

PCE† - 100A/cm2, 85oC

54% (Blue)

13% (Green)

18% (Near Red*)

7% (Amber*)

65% (Blue)

30% (Green)

45% (Near Red)

19% (Amber)

83% (Blue)

60% (Green)

70% (Near Red)

55% (Amber)

* The status of red and amber emitters is based on commercial AlInGaP LEDs. However, there is the possibility of developing InGaN or other material

system-based LEDs that emit at these wavelengths. LEDs in novel materials systems would currently have lower performance levels but may

represent the path to simultaneously meeting all the ultimate performance targets. Research on novel emitter materials is not expected to meet shorter

term performance targets but should demonstrate a clear path to meeting all 2025 performance targets.

† Optical power out divided by electrical power in for the LED package.

8

Table 3-5. Advanced Emitter Device Architectures

Advanced Emitter Device Architectures

Description: Explore the use of advanced emitter device architectures with state-of-art-emitter materials to improve

existing trade-offs between (a) the extraction of white photons from a device package, as measured by overall package

power conversion efficiency (lm/W), and (b) the ability to deliver white photons to a target, which generally improves with

luminous emittance (lm/mm2). An example of such a trade-off is droop, in which power conversion efficiency decreases but

luminous emittance increases as input current density is increased. Architectures could include the use of tunnel junctions,

photonic crystals, photonic metamaterials, stimulated emission, and/or laser devices. Of interest are both increased

luminous emittance without sacrificing power conversion efficiency, or increased power conversion efficiency without

sacrificing luminous emittance, to improve overall lighting system efficiency. Trade-offs between device power conversion

efficiency and luminous emittance (or optical distribution) should be discussed by the applicant. Device architecture

advancements would be demonstrated on blue (or possibly violet) emitters but approaches are encouraged to demonstrate

advancements in white emitting architectures as well. Proposed device architectures should enable a meaningful energy

impact.


PCE† - 35A/cm2, 25oC 67% (Blue)

[156 lm/W (Warm White)]

84% (Blue)

[237 lm/W (Warm

White)]

90% (Blue)

[249 lm/W (Warm

White)]

Luminance and optical distribution

for application efficiency

310 lm/mm2, Lambertian

distribution

500 lm/mm2, optimized

optical distribution

pattern

800 lm/mm2, optimized

optical distribution

pattern

9

Table 3-6. Understanding Quantum Dot Optical Down-Converters

Understanding Quantum Dot Optical Down-Converters

Description: Research to advance understanding in high efficiency, on-chip quantum dot (QD) down converters to match or

exceed performance of conventional on-chip phosphor materials. Research should explore QD architectures, degradation

mechanisms, synthesis techniques, and/or functionalization approaches and demonstrate advancements in on-chip LED

performance at multiple emission wavelengths relevant to high efficiency solid state lighting. Research should seek to

provide a path to a set of performance parameters that make QDs competitive with conventional phosphors for application

in general illumination. Alternatively, research could identify fundamental limitations for QD application in LED lighting

applications. Research in quantum dots that do not contain heavy metals or scarce materials is encouraged. Metrics below

describe the status of state-of-the-art phosphors used for LED lighting to provide targets for QD performance.


Quantum yield (QY) at 150°C across

the visible spectrum and at 1

W/mm2

88% (Green)

81% (Red)

91% (Green)

88% (Red)

99% (Green)

95% (Red)

Spectral FWHM 110 nm (Green)

75 nm (Red)

70 nm (Green)

30 nm (Red)

30 nm

(at all wavelengths)

On-chip reliability:

Color shift

Depreciation

Failure

∆u’v’ < 0.007

at 6,000 hours

∆u’v’ < 0.002

over life

∆u’v’ < 0.002

over life

10

Table 3-7. Advanced LED Lighting Concepts

Advanced LED Lighting Concepts

Description: Applicants can pursue one of the following two approaches.

1. Develop fully optimized color-mixed direct LED (without phosphor conversion) modules or luminaires that demonstrate

advancements in efficiency and efficacy over previous color-mixed solutions, tracking color-mixed performance shown

in Figure 3-1.

2. Develop lighting system architectures that take advantage of the unique properties of LEDs to demonstrate improved

lighting application efficiency, including advanced lighting values (e.g., human physiological benefits as demonstrated by

spectrum and intensity levels appropriate for engaging these responses). Concepts that demonstrate improvements to

lighting application efficiency should address all of the lighting application efficiency metrics described below.

Metrics 2017 Status Interim

2025 Targets 2035 Targets

Color mixed luminaire efficiency,

efficacy, and performance across

operational range

(depends on application – user may

define metrics for specific use case)

100 lm/W (3000-4000 K, 80

CRI, ANSI Quadrangle) 150 lm/W (WW and CW) 250 lm/W (WW and CW)

Lighting application efficiency

(depends on application – user may

define metrics for specific use

cases)

Luminaire efficiency: 150

lm/W


lm/W


lm/W

Task optical delivery

efficiency: depends on

application


efficiency: applicant

discuss and describe

improvement


efficiency: applicant

discuss and describe

improvement

Spectral efficiency:

depends on application Spectral efficiency: 90% Spectral efficiency: 95%

Intensity control: none or

remote at dimmer switch

Intensity control: active

and automatic

Intensity control: active

and automatic

* Spectral efficiency refers to the overlap of the emitted spectrum with the spectrum appropriate to the activity or desired visual or non-visual

response.

11

Table 3-8. LED Power and Functional Electronics

LED Power and Functional Electronics

Description: Develop advanced prototype power delivery concepts for luminaires with high efficiency, high reliability, and

minimal size and weight. Approaches should explore use of new components, devices, materials, circuits, and system designs

to provide improved performance. The integration of wide bandgap components into the driver is encouraged. Additional

advancements could include systems with multiple control channels, full dimmability, and maximum efficiency at extended

operating ranges. Size and weight advancements should demonstrate an advancement beyond existing power/weight or

power/volume relationships. Row 3 below provides a target for 100W power supplies. Work on different power levels

(higher or lower) should provide similar targets.


Power supply efficiency 88% 93% at full power

90% in dimmed state

95% at all operating

conditions

Power supply reliability

Applicant estimated

lamp/luminaire survival

factor (various methods

used)

95% survival rate with a

90% confidence level across

reported case temperature

curve

99% survival rate with a

90% confidence level

across reported case

temperature curve

Size-volume-form factor:

Lumens (or watts) per volume

(or mass)

100 W Driver:

650 g

475 cm3

100 W Driver:

300 g

275 cm3

150 W Driver:

200 g

175 cm3

Table 3-9. Additive Fabrication Technologies for Lighting

Additive Fabrication Technologies for Lighting

Description: Develop high volume additive manufacturing technologies for any portion of the LED lighting manufacturing

value chain. Approaches should be cost effective and reduce part count in the manufacturing process and be applicable to

mass production, not just prototype development. Development of printable materials with properties specific to lighting

applications is of interest for additive manufacturing approaches (optical, electronic or thermal properties). Specific

portions of processes that are of interest for additive manufacturing advancements include:

• Wafer scale packaging, including down-converter and encapsulant deposition.

• Power supply component and module manufacturing.

• Rapid creation of tooling for optics, heat sink or housing manufacturing.

• Flexible production of lighting products.

Additive manufacturing techniques apply to many different aspects of the supply chain and manufacturing processes. The

proposed approaches will need to detail the baseline performance metrics and the improvements in performance metrics

that can be obtained.

Researchers should demonstrate thorough knowledge of the portion of the manufacturing value chain they are working in

and should provide quantitative metrics, status, and targets for their research.

12

3.3 OLED Research Needs

Specific critical research tasks were suggested by members of the lighting science R&D community who work

with the BTO SSL Program. The limited number of R&D tasks reflects the practical reality that BTO must

leverage research funding for early stage research activities to achieve the most meaningful advancements

possible.

The OLED tasks identified based on stakeholder suggestions are outlined below. The milestones provided in

the tasks described below represent the minimal descriptions for progress. They provide initial and interim

targets for quantitative evaluation of progress. All these tasks will require some additional system-level

performance description and, most likely, additional metrics specific to the proposed approach. Researchers in

these areas are expected to possess and communicate a detailed, system-level understanding of the role of the

described research. Where appropriate, researchers should further define and describe metrics and milestones

that are necessary to demonstrate progress in the research topic.

3.3.1 OLED Research Tasks

Table 3-10. Stable, Efficient White Devices

Stable, Efficient White Devices

Description: Develop novel materials and structures that can help create a highly efficient, stable white device. The device

should have desirable color qualities, long lifetime, and high efficiency, even at high brightness. The approach may include

development of highly efficient blue emitter materials and hosts or, may comprise a device architecture leading to longer

lifetime, such as graded doping approaches or tandem structures with improved charge generation layers to maximize

internal quantum efficiency (IQE). Materials/structures should be demonstrated in OLED devices that are characterized to

ascertain the performance as compared to the metrics below. Novel materials/structures should demonstrate high stability,

while maintaining or improving other applicable metrics.

Metrics 2017 Status Interim 2025 Target 2035 Targets

Internal Quantum Efficiency 62% 80% 85%

Voltage per stack @ 10,000 lm/m2 2.83 V 2.75 V 2.7 V

Stability

L70: 40,000 hours

at 10,000 lm/m2 L70

Catastrophic failure

rates

Color shift

L70: 50,000 Hours at

10,000 lm/m2

L70: 50,000 Hours at

10,000 lm/m2

13

Table 3-11. Advanced Fabrication Technology for OLEDs

Advanced Fabrication Technology for OLEDs

Description: Novel approaches and advancements in materials deposition, device fabrication, or encapsulation of OLED

panels that lead to significant reduction in processing costs without unacceptably degrading performance. The techniques

must enable high yields of products that meet BTO SSL performance targets and contribute to meeting ultimate OLED cost

targets. Proposals could involve additive, maskless patterning techniques, roll-to-roll handling, high speed deposition of

organic materials or barriers, but are not necessarily limited to these topics.


Yield 60-85% 95% 98%

OLED panel production cost $100/m2 $20/m2 $10/m2

Table 3-12. Light Extraction and Utilization

Light Extraction and Utilization

Description: Devise new optical and device designs for improving OLED light extraction while retaining the thin profile and

state-of-the-art performance of OLED panels. Applicants should consider how their approach integrates and operates in

state-of-the-art structures and should include modeling or quantitative analysis that supports the proposed method.

Solutions should define a path for low-cost, scalable, and high yield manufacturing. The proposed approach can also explore

light-shaping techniques that can be integrated with the proposed light extraction technology to attain increased utilization

efficiency of the generated light. Such methods should allow some control of the angular distribution of intensity but

minimize the variation of color with angle.

Metrics 2017 Status Interim

2025 Target 2035 Target

Extraction efficiency (EQE/IQE) 55% 60% 75%

Color variation with angle (∆u’v’) < ±0.003 < ±0.002 < ±0.002

Light delivery efficiency Lambertian 20% improvement of

optical delivery efficiency

50% improvement of

optical delivery efficiency

14

Table 3-13. OLED Prototype Lighting Platforms

OLED Prototype Lighting Platforms

Description: Develop OLED lighting platforms that achieve the performance and design goals of OLED lighting technology and

demonstrate clear differentiation from existing products. The designs should embody major advances in form factor and/or light

distribution, while offering high performance and adaptability. The panels, mechanical supports, drivers and power supplies

should be consistent with innovative form factors. Advanced custom power supplies should efficiently convert line power to

acceptable input power for the OLED source(s) and maintain their performance over the life of the device. Innovations may

include but are not limited to: unique form factor (thin, flexible); beam control; spectral tunability; modularity; automated

control of intensity and color in aging. Proposals should provide quantitative targets for distinctive performance.

Metrics 2017 Status Interim 2025 Target 2035 Target

Light engine

Power supply efficiency 50-80% 90% 95% at all operating

conditions

Driver form factor Typical driver box Thin to match design of

luminaire

Unobtrusive power

supply integration

Light engine efficacy N/A 135 lm/W 165 lm/W

Luminaire Efficacy 20-50 lm/W 120 lm/W 150 lm/W

15

3.4 Cross-Cutting Lighting R&D (LED and OLED)

Specific critical R&D tasks were suggested by members of the lighting science R&D community. The limited

number of R&D tasks reflects the practical reality that BTO must leverage research funding for early stage

research activities to achieve the most meaningful advancements possible. The topic of understanding

physiological impacts of light was consistently prioritized by the BTO SSL Program stakeholders as critical to

advance solid-state lighting and to secure its energy-saving potential. Uncertainty around physiological

responses to light can slow adoption of energy-saving SSL, and uncertain effectiveness of visual and non-

visual physiological responses to light can reduce the efficacy of the lighting system. Improved understanding

of this topic will encourage engineering and technology advancements. A more detailed discussion of this

topic is found in Section 4.

The milestones provided in the tasks described below represent the minimal descriptions for progress. They

provide initial and interim targets for quantitative evaluation of progress. All these tasks will require some

additional system-level performance description and, most likely, additional metrics specific to the proposed

approach. Researchers in these areas are expected to possess and communicate a detailed, system-level

understanding of the role of the described research. Where appropriate, researchers should further define and

describe metrics and milestones that are necessary to demonstrate progress in the research topic.

Table 3-14. Understanding Human Physiological Impacts of Light

Understanding Human Physiological Impacts of Light

Description: Research to understand and define physiologically optimized lighting for the general population based on

objective physiological responses to light or large-scale collection or review of subjective responses. Specific aspects to

understand could be optimum and threshold intensity, duration, and spectrum for light during the day and pre-sleep.

Specific R&D could be performed on sub-populations that could inform guidance for the general population. R&D efforts

should advance lab-scale studies to more naturalistic studies that can guide development and implementation of lighting for

positive physiological responses.


Human physiological impacts

Lab studies in unrealistic

lighted environments; or

subjective response with

limited participants

Understand lighting

thresholds in realistic

settings for physiological

responses across all

types of lighting

applications to maximize

efficiency and safety

Broad implementation of

efficient lighting that

reduces or eliminates

negative physiological

impacts of lighting

16

Table 3-15. Understanding Lighting Application Efficiency

Understanding Lighting Application Efficiency

Description: Develop a general framework, mathematical model, and computer simulation approach to characterize lighting

application efficiency for any lighting application in terms of the four primary aspects of lighting application efficiency: light

source efficiency, optical delivery efficiency, spectral efficiency, and intensity efficacy. Light source efficiency describes the

efficiency of the lighting product in generating light from input electrical watts. Optical delivery efficiency describes how

efficiently light is delivered for all of the various ‘jobs’ of the lighting. Spectral efficiency defines the overlap of the ultimate

spectrum that reaches the task or eye with an optimum spectrum for the activity or intent of the lighting, e.g. visual acuity,

color rendition, engagement of physiological responses, etc. Intensity efficacy describes the difference between the intensity

of the provided light and the optimum intensity for the specific intent of the light. Optical delivery, spectral efficiency, and

intensity efficacy may have temporal dependency as occupant positioning and activities in a space change over time. The

proposed R&D and resulting models should be validated with lighting mock-ups with optimized light placements and optical

distributions and then measured.

Project status and metrics for progress for this R&D task should be supplied by researchers in this topic. The near term

objective for this R&D task is to develop a working framework and vocabulary to characterize Lighting Application Efficiency

in any lighting application. The framework should allow accurate computer modeling of Lighting Application Efficiency and

this should be validated in the research against real lighting situations. In addition, the research should provide initial

characterization of Light Source Efficiency (this should be readily available), Efficiency of lighting delivery to receptor

(typically the eye), Spectral efficiency, and Intensity effectiveness.


Lighting Application Efficiency

framework and model

No comprehensive

framework or model

Application agnostic

model that can be used

to optimize total Lighting

Application Efficiency

Ubiquitous use of

Lighting Application

Efficiency modeling for

building, room, lighting

layout, and product

design

17

Directions in Lighting Science SSL offers vast opportunity to improve the efficiency, performance, and value of lighting by creating new

applications and benefits. The initial motivations for the pursuit of light-emitting diode (LED) and organic

light-emitting diode (OLED) technology were the promise of high-source efficacy and the prospects of

leveraging semiconductor manufacturing processes. While there is still considerable room for improvement,

SSL is already fulfilling its promise as it continues to demonstrate improved efficacy over conventional

lighting sources, long lifetimes that enable payback within reasonable time periods, and new features including

spectral tunability, advanced controls, and connectivity. These attributes have led to rapid adoption of SSL,

already resulting in significant energy savings. In addition, as the technology has developed, it has become

clear that the impacts of SSL will go far beyond energy savings alone. SSL has the potential to have profound

beneficial impacts on the environment, horticulture, transportation safety, human health, and productivity, all

of which can be realized while saving significant amounts of energy compared to conventional lighting

technologies.

SSL holds the promises of ongoing energy savings and new lighting value, but continued R&D is required to

fully realize these promises. In particular, the emergence of SSL has enabled key lighting science discoveries

that reveal many new and essential connections between lighting exposure and biological impacts. These links

further indicate that existing performance parameters, metrics, design norms and guidance are insufficient for

describing how lighting should be specified for an application. R&D into lighting science is necessary to

ensure that SSL continues to drive energy savings, since the growing use of color tuning, controls and complex

sensor networks could result in an increase in energy consumption without targeted research. The following

sections explore many of the key directions in lighting science for SSL.

For the past 15 years, progress in SSL has been measured in lumens per watt (lm/W). This framework

describes the productive aspect of lighting, the lumen, in the numerator and consumption aspect of producing

light – the input electrical power required – in the denominator. Though this metric is easy to use, additional

information is required to ensure that the lumens are effective for the intended lighting application. Typically,

additional color quality metrics, including color rendering index (CRI) and correlated color temperature

(CCT), are also provided to aid in the selection of an application-appropriate lighting product.

While the lumen measures the amount of light generated, it is dependent on the human response to light

represented by the action spectrum of the photopic eye response curve. As shown in Figure 4-1, the eye is

more sensitive to light in the green spectral regions, so those photons get weighted higher than photons in the

deep red or violet spectral regions. Lumens are then calculated from the convolution of the emitted spectrum

from a light source and the photopic eye response curve. Therefore, light emission that falls outside of the

photopic eye response curve shown below in Figure 4-1 is weighted less and does not contribute as strongly to

perceived lumens.

18

Figure 4-1. Human Eye Sensitivity Function CIE 1978 Photopic Vision

Source: LEDs Magazine, Radiometric and Photometric Terms [3]

This lumen framework represents the foundation of the commonplace efficacy metric, as measured by lm/W,

which has been effective in creating efficient SSL solutions and driving performance improvements to date.

However, the improvements in LED technology have brought new levels of controllability and functionality,

which leaves the metric lm/W unable to fully describe the efficiency, effectiveness, and value of LEDs in

many lighting applications. Two examples where the lm/W metric falls short are human non-visual

physiological responses to light and horticultural lighting. As discussed further in Section 4.2, both of these

applications use different wavelengths of light to impact physiological activity, also known as action spectra,

to define the effectiveness of lighting.

In new and pre-existing lighting applications, the effectiveness of lighting may not be solely described in terms

of lumens. New action spectra will be used to describe other aspects of lighting effectiveness beyond human

visibility. With the new levels of spectral control offered by SSL, the overlap of emitted light with new action

spectrum can be optimized. In addition, characterizing lighting systems only in terms of emitted lumens does

not account for downstream elements of the lighting system that can greatly affect efficiency. SSL

technologies offer precise control of the intensity and direction of emitted light. This control can enable more

efficient delivery of just the right amount of light to the right place at the right time. Taken together, precise

spectral, optical, and intensity control enable a new frontier of energy savings that is enabled by the high

starting efficiency of the SSL source.

While imperfect to describe the full performance in a light application, the lumen still remains the primary

metric for describing lighting service since the primary function of light will be to illuminate the physical

space and objects within it. However, developing a more holistic framework for characterizing light that

reaches the intended target (whether that target is biological or an inanimate object) for the intended

application will be beneficial to demonstrate all of the improved functionality of SSL technology.

19

4.1 Lighting Application Efficiency – A New Framework and Opportunity for Energy

Savings

A new framework for characterizing the effectiveness and efficiency of a lighting system would improve the

way we differentiate lighting performance for a given application. The new framework would potentially need

to consider the:

• Spectral efficiency (described in Section 4.1.1);

• Light source efficiency of the luminaire (described in Section 4.1.2);

• Optical delivery efficiency (described in Section 4.1.3); and

• Intensity effectiveness (described in Section 4.1.4).

Combined, these four elements could be used to describe the overall lighting application efficiency

demonstrated with SSL systems. While each has been demonstrated, evaluated, and studied independently with

SSL, they largely have not been considered holistically within a common framework. If each component of

lighting application efficiency continues to be evaluated in isolation, this will inherently limit their use within

industry and represents a missed opportunity for energy savings. A holistic framework would enable the

different aspects of lighting application efficiency to be considered and optimized for different applications.

This proposed framework, shown in Figure 4-2, would also guide future R&D in lighting application

efficiency to target the most impactful aspects of performance for a given application.

Figure 4-2. Possible Framework for Evaluating Lighting Application Efficiency

The concept of a holistic framework for considering lighting application efficiency is new, particularly when

considering the new capabilities of SSL technology. As such, the definitions and framework described here

and in the following sections are preliminary.

4.1.1 Spectral Efficiency

Spectral engineering has been a central theme of SSL since its beginning – with significant effort and research

towards improving the lighting performance as measured by many of the most common metrics such as

lumens, CCT and CRI. In terms of spectral optimization, this is just the beginning, and many applications will

benefit from more finely controlled spectra, not only for reducing the energy required for the application, but

also for improving occupant well-being and productivity. Spectral efficiency can be defined as the ability of

the emitted spectrum to produce the desired response for a given application. Using a tailored spectrum for an

application enables maximization of desired wavelengths and the omission or reduction of damaging or

unnecessary portions of the spectrum for that task. This concept is still relatively novel for general illumination

applications, since fine spectral control has only been made feasible by the recent onset of efficient LED

technology. As such, for most lighting applications, the optimum spectrum is not well understood, and

significant research will be necessary to develop this understanding.

20

Spectral efficiency describes how well the spectral power density (SPD) emitted by the lighting product

overlaps with the desired action spectrum for the intent of the lighting application. Currently, almost all

lighting products are designed for overlap with the photopic eye response action spectrum (which gives rise to

the definition of the ‘lumen’) shown in Figure 4-3. In addition, future lighting systems will need to consider

additional action spectrum beyond just the eye response. As research on the physiological responses to light

has progressed, action spectra have been developed for physiological responses such as melanopic response,

reflective contrast (visibility of colors), color saturation, plant growth response, and animal responses5, as

illustrated in Figure 4-3 below. Future lighting systems must consider different or even multiple action spectra

and the ultimate spectral effectiveness of these systems would be characterized by a spectral effectiveness per

watt term, similarly to lumens per watt. To develop understanding in physiological responses and the possible

benefits from advanced lighting, it is critical to collaborate with a broader set of stakeholders represented by

federal agencies, such as NIH, USDA, and more.

Figure 4-3. Action Spectra for Humans and Plants: (a) Human Action Spectra [4] (b) Plant Action Spectra [5]

4.1.2 Source Efficiency

With the expanded range of possible lighting functions with different action spectra discussed in Section 4.1.1,

the properties of the spectrum must be reported precisely. The use of radiometric efficiency and spectral power

density will allow the precision of describing the tuned spectra in ways that CCT and CRI never could. These

parameters can then be used to determine spectral efficiency for the application – i.e., how well the emitted

5 Animal and plant responses to light are not identified as priority R&D topics by SSL Program stakeholders, but spectral responses for these

applications exemplify additional benefits to spectral control offered by solid-state lighting technology.

21

spectrum overlap with the action spectrum for the given application. In some cases, multiple action spectra and

associated metrics of effectiveness may need to be considered for the lighting application.

Currently, an average performing LED lighting product (including power supply) is approximately 42%

efficient at converting input electrical power to light, also described as power conversion efficiency (PCE).6

Discussions for improvements to the source efficiency are covered in detail in Section 5 and Section 6 of this

document. With the advancements in the power conversion efficiency of SSL components in recent years, the

focus can now include other elements of lighting application efficiency. However, there are still trade-offs

between source efficiency and spectral efficiency. For example, high CRI sources have lower source efficiency

due to additional phosphor conversion losses. For applications that need a high CRI source, it is appropriate to

trade-off efficacy (reduction in spectral overlap with eye response action spectrum) for color quality, although

work is ongoing to minimize this trade-off. Designing lights to other action spectrum might involve different

trade-offs between source efficiency and spectral efficiency. Similarly, optical delivery efficiency may involve

trade-offs between source efficiency and high luminance (see Section 5.4.2), which is typically required for

advanced optical control. Finally, in order to achieve optimum light intensity efficiency, it may be necessary to

dim the light source, though many integrated lighting products suffer from reduced power supply efficiency as

the light is operated at dimmed settings. This is another trade-off that can be considered within the lighting

application efficiency framework. Understanding the trade-offs in efficiency losses between source efficiency

and the other elements of lighting application efficiency can guide lighting system design and R&D to reduce

these trade-offs.

4.1.3 Optical Delivery Efficiency

LED lighting luminaires can be very efficient at producing light with efficacies reaching 200 lm/W. While

improving efficacy is important, the effective delivery of the light to the target is another way to improve

lighting efficiency. Additional research is needed into optical control of lighting to put more of the light

generated to use in the application. The definition of ‘useful light’ for the application will vary and also needs

additional research to define what is “useful.” In some cases, it is light on the horizontal work surface, in other

cases it is putting light on a specific object and, ultimately, it is delivering light to the eye for both visual and

non-visual physiological effects.

While large amounts of lumens are emitted, the proportion of light emitted from luminaires in a room that

reaches the eye range is estimated to be 1x10-6 to 1x10-9 [6]. 7 This is highly dependent on the room geometry,

lighting layout, occupant orientation, reflectance of surfaces, and more, but these estimates indicate that there

is room for practical improvement in optical delivery efficiency without greatly affecting the lighting scheme.

Lighting layouts are currently designed to meet even illuminance levels on horizontal surfaces in a room,

which inherently does not account for the light levels reaching the vertical orientation of the human eye, nor

the illumination required for foveal and peripheral vision. While lighting design choices are complex and

consider a variety of additional characteristics such as product cost and aesthetics, it is important to consider

the value of having metrics that would more accurately specify the lighting requirements for occupants.

However, when considering the pathways to achieving greater optical delivery efficiency, there are significant

barriers. For example, retrofit LED lighting products are constrained by performance limitations, luminaire

layouts, and optical distributions of traditional lighting technology and are not optimized to effectively provide

light that reaches occupants’ eyes. In general, lighting products are not designed to precisely deliver the

necessary amount and type of light for the intent of the application, and instead are largely focused on

replacing legacy lighting systems. Breaking from this paradigm will require advancements in lighting design

practices and advanced software tools as well as improved lighting products with controllable optical

distribution.

6 This analysis considers a warm white lighting product (3000K, CRI 80) with an efficacy of 137 lm/W and a luminous efficacy of radiation (LER) of

323 lm/W as described in the 2017 DOE SSL Suggested Research Topics Supplement in Table 4.2. Dividing the actual efficacy of a lighting product

by the LER gives the power conversion efficiency for the product. 7 Estimates based on: 1) geometrical analysis of the human field of vision within a typical room, 2) unpublished calculation of measured light received

at eye compared to total emitted light in a specific test room.

22

Simply doubling the optical delivery of light to the eye is a conservative target that will require significant

R&D. In particular, research is needed to develop an approach to broadly understand and characterize the

status of optical delivery efficiency for all lighting applications for both visual perception of the space and the

non-visual biological factors. Once this foundational knowledge is established, effort can turn to the

development of software for the optimization of lighting layouts and product optical distributions. In addition

to knowing the optimal optical and spectral distribution of the light for various lighting applications, the ability

to put the light where it is needed must be improved. Research into new optical control technologies in

luminaires systems to improve the efficiency and directionality in primary and secondary optics is required.

Some of the various technologies to be investigated include metamaterials, new device structures, and tunable

optics. Furthermore, efforts are then needed to inform revised lighting design practices as well as the

deployment of new software tools and optical delivery approaches.

4.1.4 Intensity Effectiveness

Conservatively, 50% (or more) of generated light is produced when there is no observer present to see the

light, both in buildings and on roadways, leading to wasted light and energy [6]. This inefficient use of light

can be improved with better intensity controls. Intensity effectiveness is a term to ensure only the right amount

of light is used at a given time (when observers are in the lighted area), and also to help characterize when

insufficient light intensity is being provided, impacting the effectiveness of the intent of the light. LED lighting

can usher in better intensity effectiveness since it is inherently controllable compared to traditional lighting

technologies with its full, instantaneous dimmability. For example, when there is sufficient daylighting or light

is not needed, products can be dimmed or turned off to save energy.

Controls and sensors are commercially available and have been deployed to improve intensity effectiveness.

However, it is necessary to improve performance, cost, and consumer confidence to increase adoption of

controls, and hence, intensity effectiveness. In addition, controls and SSL sources need to be efficient and

consume little power in their dimmed and standby- or off-states so that energy savings are not overshadowed.

In terms of lighting science, new guidance needs to be developed for the optimum intensity levels for basic

illumination of objects and also for engaging non-visual physiological responses. In addition, it is important

that future industry guidance also focus on the color and intensity of light that reaches the eye, not just what is

delivered to a surface. Different lighting applications, from roadway to office to industrial, etc. will offer

different prospects for understanding the optimum lighting intensity and for engaging controls to get the

intensity right. Also, as with the other elements of lighting application efficiency, there may be conflicting

demands on the lighting system. Light levels for illumination and performance of tasks may be different than

optimum light levels for physiological responses. In a lighted space there may also be different intensity

requirements for different population segments. Older eyes require higher illumination levels. Many of these

considerations are well known to skilled lighting practitioners, but SSL technology provides new levels of

control of the light intensity that were not previously practical. A framework for considering the various

intensity requirements and possibilities for active control in a space could guide lighting designers toward the

most efficient and practical solutions for optimizing the light intensity levels.

4.2 Visual and Non-Visual Responses to Light

Humans are continuously exposed to natural and electric lighting, all of which has some effect on our

physiology, regardless of the source. It is now clear that light has important effects that can be harnessed for

improved health and well-being. Recent research has advanced the understanding that light not only enables

vision, but also is a critical signal to our biological systems, affecting circadian rhythms, pupillary response,

alertness, and more.

Circadian rhythms are inherently tied to the natural daylight cycle – light enhances wakefulness and darkness

promotes sleep. In the natural environment, humans would rise in the morning when the sun comes up and

emits high intensity, blue-enriched light. Humans then begin to get tired in the evening when the natural light

is blue depleted. With the advent of electric lighting, natural cycles have shifted. Electric lighting not only

enables sufficient light to work indoors during the day, it also extends the natural day and provides different

23

lighting signals than the natural environment. Conventional lighting technologies and practices provide

relatively low intensity levels during the day compared to daytime sunlight conditions (in most places). On top

of this, most indoor lighting has a relatively ‘warm’ color temperature (lower CCT) than natural daylight

conditions. In the evening, pre-sleep, there may be too much light with too much blue content, particularly

when considering exposure to electronic displays (TVs, tablets, etc.). These electric lighting signals may

confuse our natural diurnal rhythms. This basic understanding of these physiological responses to light is new

and coincides with the development of SSL technology. SSL technology enables lighting systems that can be

designed to engage with human physiological responses; however, clear guidance is needed so that light

sources can provide physiologically effective light without compromising efficiency.

4.2.1 R&D Directions for Visual and Non-Visual Responses to Light

The objective of the BTO SSL Program is to reduce energy used for lighting. However, R&D stakeholders

have identified the topic of eye-mediated physiological responses to light as critical for the advancement of

SSL technology and adoption. It is now clear that existing lighting technologies and solutions may not be

optimized for health and well-being. Light levels during the day may be too low with insufficient blue content.

And light levels at night, particularly indoors, may be too bright with too much blue content – which

counteracts the human lighting needs for natural circadian regulation as described in the previous Section 4.2.

Lighting is a ubiquitous part of the built environment and should be designed and deployed first for optimum

health and well-being, and second for high efficiency. Therefore, it is critical that physiological responses to

light are clearly understood and then efficiently acted upon, in other words, light sources need to first be

effective and then be efficient for illumination and optimum health. The BTO SSL Program is developing

agreements to collaborate closely with the National Institutes of Health in this area.

While basic guidance is available, the current knowledge base is limited and much of the underlying

physiological responses to light are still unknown or being validated in clinical trials. More R&D is needed to

fully understand what lighting inputs produce the different non-visual effects, and how to properly control

them in a real-world setting. Continuing fundamental science research is needed to understand and map the

physiology responses to lighting. In addition, more research is needed to validate and translate the findings of

basic, lab-scale research to realistic lighting use cases. Only then can we begin to guide lighting technology

development and lighting design to elicit these physiological responses in an energy efficient manner. In

particular, it is important to better understand dose-intensity-color relationships for light and physiological

responses in realistic lighting contexts to guide lighting practice. This understanding will guide lighting

systems to be effective at engaging physiological responses, which, in turn, will enable lighting manufacturers

to develop the most efficient solutions. Some specific directions for research in the topic of physiological

responses to light were provided by R&D stakeholders to the BTO SSL Program, listed below:

1. Understanding fundamental non-visual physiological-light interactions, including endocrine responses

and other eye-mediated, physiological responses.

2. Understanding dose-intensity-color relationships and physiological threshold levels. At the high end,

where more light or more blue-enriched light dosing does not increase alertness or melatonin

suppression. At the low end, where light levels are too low to affect melatonin secretion.

3. Understanding synergistic interactions between different colors of light. Does white with blue-

enriched content have the same physiological impact as monochromatic blue light with same blue

content and how do metamers with different spectral content but the same color point affect

physiological responses?

4. Develop best practices for measuring the physiological responses of occupants in a realistic lighting

context.

5. Develop a process to determine when published scientific evidence is sufficient to define the

physiological response to light in realistic lighting context and consensus is reached to begin to

develop lighting guidelines based on the evidence.

As research findings are published there needs to be a robust mechanism for evaluating the findings, and if

corroborated, transitioning the findings to standards bodies who can provide guidelines and best practices to

24

the lighting industry. There should also be a means to identify and question lighting product claims that are not

supported by academia and scientific evidence as shown in the published research results.

4.2.2 Understanding Relationship Between Energy Savings and Wellness Implications

As stated above, effectiveness of lighting systems for illumination and health must be the primary

consideration, and then efficiency. SSL offers improvements to lighting performance and human health and

well-being, while also enabling substantial energy savings. Recent R&D into physiological responses to light

is providing guidance as to how lighting systems can be optimized for human health and well-being, though

additional R&D is required to understand the energy consumption implications for possible new lighting

guidelines.

Early guidance suggests that higher light levels, particularly in the morning (but also possibly throughout the

day) will have an alerting effect on humans and improve melatonin function at the end of the day. However, it

is important to recognize that higher light levels will require increased energy consumption. Research is still

ongoing and has not yet generated guidance as to how high the light levels need to be, how long they need to

be high, and how much blue content in the light is necessary. More R&D is needed to understand the

relationships between timing, duration, intensity, and color of light, as well as the threshold where positive

effects saturate. This understanding would enable lighting solutions that maximize both effectiveness and

efficiency for daytime lighting environments.

In some lit spaces, researchers are investigating the benefits of lighting that can be tuned from higher blue

content white during the day, to lower blue content light in the evening. This dynamic controllable light would

provide a signal to our body throughout the day and is thought to improve circadian regulation and appropriate

melatonin secretion. However, this approach affects energy consumption because tunable lighting products are

generally not as efficient as static non-tunable lights. Tunable products consume more energy because they

require multiple types of LED sources that may not be operated at their peak operating range. Therefore, the

resulting white light from multiple color mixed LED sources is often less efficient than static non-tunable

white LED sources that employ white LEDs. These systems could also have increased optical losses due to the

mixing of different color LED sources. And, multi-channel power supplies are necessary to control the

different color LED sources, which could result in non-optimized operating conditions. The technical

efficiency reductions in tunable lighting products are not fundamental to LED technology, but they do present

significant engineering challenges for the lighting product design. These considerations for high efficiency

operation of the LED sources and power supplies at a broader range of operating conditions have greatly

influenced the priority R&D tasks described in Section 5.

The current understanding of how lighting can be optimized for health is limited, and therefore we do not yet

know the full extent of the lighting energy consumption impacts. However, it is clear that increased R&D

investigating the physiological responses in common lighting applications, in conjunction with R&D in the

engineered color-tunable lighting systems, will enable the development of products that can most effectively

and efficiently provide lighting for human health and well-being.

4.3 Connected Lighting

The replacement of the lighting infrastructure with LED products offers the potential for future connected

lighting systems (CLS) that could become a platform that enables greater energy savings, lighting effectiveness

for new lighting applications, and high-value data collection in buildings and cities. For example, as lighting

systems become more connected, it is anticipated that they will increasingly offer the ability to optimize

resources and processes, deliver health and productivity gains, and yield new revenue streams. Further, it is

likely that these capabilities will offer benefits that match or exceed the value of the energy savings they

deliver. The value of services made possible by data from networked SSL systems might partly or fully offset

the incremental costs of sensors, network interfaces, and other additional components. Systems made up of

connected lighting devices could become data collection platforms that enable even greater lighting and non-

lighting energy savings in buildings and cities, and much more.

25

4.3.1 Energy Savings and Other Valued Features

As SSL technology matures, maximizing the energy savings from connected SSL systems will become

increasingly dependent on successful integration into the built environment. Lighting controls have the

potential to deliver significant energy savings by adjusting the amount and type of light to the real-time needs

of a particular space and its occupants. SSL products are poised to be the catalyst that unlocks the energy

savings potential of lighting controls due to their unprecedented controllability and increasing degrees of

automated configuration – facilitated by embedded sensors and intelligence, as well as by other features and

capabilities that leverage the data they collect. Lighting systems that can leverage occupancy sensing, daylight

harvesting, high-output trim, personal area controls, or any combination of these approaches have been shown

to provide energy savings of as much as 20% to 60% of SSL power consumption, depending on the application

and use-case [7].

The ability of connected lighting to collect and exchange useful data, and possibly even serve as a backbone of

the fast-emerging Internet of Things (IoT), offers the potential to enable a wide array of services, benefits, and

revenue streams that enhance the value of lighting systems and bring that improvement to building systems

that have long operated in isolation. Connected lighting systems can help building owners to understand how a

space is being utilized by its occupants and to deploy adaptive lighting strategies that increase lighting energy

efficiency. A lighting-based advanced sensor network can provide a vast array of data from the building

environment (e.g., energy usage, temperature, daylighting) or building activity (e.g., occupancy, asset location

and movement). This information can be used to improve energy savings through daylight harvesting,

occupancy detection, demand response programs, time-of-day dimming schedule, and real-time energy savings

reporting. Other information can lead to better utilization and maintenance of the building including advanced

occupancy detection, light-level stability, personalized setting profile, and fixture outage reporting.

In addition to a range of occupancy and daylight sensors, other types of sensors could be installed, including

those to measure carbon dioxide, imaging, vibration, sound, and barometric pressure — resulting in such

“smart city” features as air quality monitoring, weather warnings, theft detection, guidance to available parking

spaces, and transit optimization. Connected street lighting systems offer the ability for city officials to

implement adaptive lighting strategies (e.g., having the street light at 100% brightness when it turns dark and

gradually dim to 50% in the middle of the night and return to full brightness in the early morning for

commuters) that deliver further energy savings. Connected street lights may also provide the city the location

of each light pole to better manage these assets, particularly when there are failures.

If connected lighting products have the capability to self-measure and report energy use, utilities could offer

incentives to customers based on actual savings instead of estimated savings. Data-driven energy management

can significantly reduce energy consumption and enable new market opportunities, such as pay-for-

performance energy efficiency initiatives; energy billing for devices currently under flat-rate tariffs; verified

delivery of utility-incented energy transactions (e.g., peak and other demand response); lower-cost, more-

accurate energy-savings validation for service-based business models; and self-characterization of available

(i.e., marketable) “building energy services.”

SSL is already being used as a platform for indoor positioning in retail and other heavy-traffic buildings, by

using Bluetooth® and/or visible light communication to provide personalized location-based services for

occupants via mobile devices. Retailers use the luminaires to transmit to shoppers location-specific data such

as discount coupons or where in the store to find products. Beacons embedded in LED luminaires allow for the

monitoring and analysis of building use and traffic, which can lead to operational efficiencies, enhanced safety,

and increased revenues in spaces such as airports, shopping malls, logistics centers, universities, and healthcare

facilities. Connected lighting is also being considered as a promising new source of broadband communication

called Li-Fi, which modulates light to transmit data. Additionally, connected lighting is being combined with

spectral tuning in a variety of settings, with the goal of engaging physiological responses to improve mood,

productivity, and health.

26

4.3.2 Lighting Controls Interoperability

Just as SSL technology brought many new players (e.g., semiconductor manufacturers and microelectronic

system developers) to the lighting industry, the coming intersection of lighting, communication networks, big

data, and advanced analytics – facilitated by the IoT – will significantly alter the lighting industry landscape.

CLS will need to operate within the larger environment of building energy management technologies. The

challenge is agreeing on common platforms and protocols, among lighting products and within the larger IoT

landscape, which will unlock the full potential of IoT by enabling the exchange of useable data among lighting

systems, other building and control systems, and the cloud. Interoperability is considered to be the pivotal

enabler of and catalyst for IoT deployment and, thus, CLS adoption and associated energy savings [8] [9].

Enabling the right level of interoperability is crucial for devices, applications, networks, and systems to work

together reliably and to securely exchange data.

Traditionally, there has been little-to-no interoperability between competing lighting-control devices and

systems, as manufacturers have focused on developing and promoting proprietary technologies or their own

version of industry standards. The benefit of interoperability is that it enables different devices, applications,

networks, and systems to work together and exchange data. For users, it reduces the risk of device or

manufacturer obsolescence, as well as the risk of having limited hardware, software, data, and service choices.

It also improves system performance by facilitating multi-vendor systems, reducing the cost of incremental

enhancement, enabling greater data exchange, and encouraging service-based architecture.

Interoperability requires industry to agree on common platforms and protocols that enable the transfer of

usable data between lighting devices, other systems, and the cloud. A number of consortia are working to

establish common specifications and standards that support increased interoperability, including the Open

Connectivity Foundation, the TALQ Consortium, oneM2M, Bluetooth special interest group, the Industrial

Internet Consortium, and the Zigbee® Alliance. As with the development of computing technologies, these

groups are taking different approaches or addressing different parts of the puzzle. There is currently little

interoperability among commercially available connected lighting systems.

4.3.3 Connected Lighting Test Bed

The BTO SSL Program is working closely with industry to identify and collaboratively address the technology

development needs of connected lighting systems. Central to the BTO efforts is a connected lighting test bed

(CLTB), designed and operated by Pacific Northwest National Laboratory (PNNL) to characterize the

capabilities of connected lighting systems. The results of these studies will increase visibility and transparency

on the capabilities and performance of new devices and systems and create information feedback loops to

inform technology developers of needed improvements as it relates to DOE priority areas of energy reporting,

interoperability, configuration complexity, cybersecurity, and key new features.

The CLTB has infrastructure that enables the efficient installation of indoor and outdoor lighting devices. Two

ceiling grids are available for installing indoor lighting luminaires. The height of each is vertically adjustable,

to enable easy installation and set varying luminaire heights. The grids have plug-and-socket interfaces to

enable easy electrical connections, and circuit-level power and energy metering in the electrical panels that

serve them. The CLTB also has dedicated infrastructure for street lighting luminaires; again, plug-and-socket

interfaces enable easy electrical connections.

To enable the testing of multiple devices and systems, the CLTB includes a software interoperability platform

that allows installed lighting devices and systems not natively capable of exchanging data with each other to be

able to communicate. Multiple commercially available indoor and outdoor connected lighting systems have

been installed in the CLTB, incorporated into the software interoperability platform, and made available for

connected lighting systems and other studies.

The CLTB is being used to investigate several areas and capabilities of connected lighting systems including

interoperability, energy reporting studies, and cybersecurity testing. A recent study focused on interoperability

as realized by the use of application programming interfaces (APIs) in several connected lighting systems and

27

characterized the extent of interoperability that they provide [10]. The APIs provided by current market-

available CLS vendors can be utilized to facilitate some interoperability between lighting systems which

enables lighting-system owners and operators to implement a basic level of multi-vendor integration and

remote configuration and management services, as well as some adaptive lighting strategies. However, in

many instances, API inconsistency and immaturity unnecessarily increase the effort required to implement

these services and strategies and reduce the value and performance that they deliver. API developers should

explore and attempt to implement common approaches to naming and organizing resources, as well as

common information and data models – which are key to both minimizing the effort required to integrate

heterogeneous systems and enabling functional, high-value use-cases.

4.3.4 Security

As more devices are becoming part of a connected world, the benefits come with security risks. This has been

demonstrated by a few publicized cases in which firewalls have been breached by hacking into lighting

products [11]. An Internet-connected lighting system can provide hackers entry points to everything behind the

network firewall, e.g., a home computer, a retailer’s payment terminals, or a government office’s sensitive

database. Studies found that even the most basic security practices that could have prevented these breaches

were often not followed, including the lack of encryption and authentication, the use of clear-text protocols to

transmit sensitive information (e.g., passwords), and the use of default passwords in customer environments

[11]. Because of these potential vulnerabilities, it is imperative that manufacturers integrate security into their

product and software development lifecycle right from the start.

Connected lighting systems and other IoT systems require further work in integrating end-to-end security.

Lighting fixtures must have authentication and security certificates for each node and the sensor data needs to

be “signed” to make sure it is coming from the correct sensor. In many cases, IoT systems will not be a single-

use, single-ownership solution. The devices and the control platform where data may be collected and

delivered can have different ownership, policy, managerial, and connectivity domains. Consequently, devices

may be required to provide access to several data consumers and controllers, while still maintaining privacy of

data where required among those consumers. Information availability with simultaneous data isolation among

common customers is critical. Securing user data and privacy, ensuring availability, and protecting network-

connected devices against unauthorized access will be crucial to companies wanting to gain and maintain trust

with connected lighting buyers.

DOE is collaborating with Underwriters Laboratory (UL) and other Industrial Internet Consortium (IIC)

members of a Security Claims Evaluation Testbed on their efforts to develop test methods for cybersecurity

vulnerabilities. Evaluation of a recently completed V0 test method is under way in the DOE CLTB using a test

setup that currently comprises a cybersecurity gateway, two commercially available cybersecurity software

services, and a Kali Linux system. When one or more test methods are deemed sufficient, DOE will conduct

studies to evaluate the cybersecurity vulnerabilities in connected lighting, and perhaps the effectiveness of

strategies and technologies for addressing them.

28

Directions in LED Science LED lighting technology has improved dramatically over the past decade to achieve among the highest

efficiencies of available white light sources. Improvements in manufacturing has enabled LED products to

achieve a low enough cost to drive measurable LED adoption in all general illumination applications. Despite

this progress, further improvements are possible and necessary to ensure even more energy savings. LED

lighting efficiency and other features, such as color quality, light distribution, form factor, and architectural

integration, have room for further advancements. The manufacturing technology for LED lighting also can be

improved to reduce cost and increase market penetration, resulting in the greatest possible energy savings for

the nation.

The following sections explore the current status, performance improvement opportunities, and challenges for

LED technology identified in Section 3.2. The key challenges currently facing LED technology also represent

some of the greatest opportunities for performance gains. The sections cover both the LED package, which

creates the white light, and the LED luminaire, which houses the LED package and provides the appropriate

interface between the electrical supply, mechanical integration, thermal handling, and optical distribution.

5.1 White LED Technology

Two common architectures for generating white light will be the focus for the discussion in the following

sections; the phosphor-converted (pc) LED based on a blue LED pumping yellow and red wavelength optical

down-converters (typically phosphors) to produce white light; and the color-mixed LED (cm-LED) approach

using primary colors that compose a red, green, blue, and amber (RGBA) LED combined to produce white

light. These are illustrated below in Figure 5-1, with the corresponding optical spectral distributions of these

white LED architectures shown in Figure 5-2.

Figure 5-1. Schematic of Two Main White LED Architectures

Note: (a) the phosphor-converted (PC) LED using blue LEDs to pump yellow and red down-converters; (b) the color-mixed

(CM) LED using direct emission LEDs to provide the different colors to mix to white.

(a) (b)

White Light

Blue LEDs + Phosphor

White Light

Direct Emission LEDs

PC-LED CM-LED

29

Figure 5-2. Typical Simulated Spectral Power Density for White-Light LED Package Architectures

Note: In all cases, the peak wavelengths and relative intensities are those which maximize LER for a 3000K CCT (warm

white), a “standard” CRI Ra of 80 and a CRI associated with the ninth, deep-red Munsell color sample R9 >0. The spectral

widths of the various source colors correspond to the current state-of-the-art. Overlaid on each spectrum is the spectrum

from an incandescent blackbody source at 3000K.

The pc-LED architecture is by far the dominant white light architecture. It has three major advantages:

simplicity (only one LED type), temperature robustness (the InGaN blue LED and YAG phosphor down-

converters can operate at relatively high temperatures), and color stability (the fractions of red, green, and blue

source colors are determined during manufacture by the phosphor optical density and are relatively stable over

time). Figure 5-3 shows a history of the luminous efficacy of pc-LEDs since the BTO SSL Program began and

the progress that has been made. It is important to note that the assumed operating conditions for qualified data

points may not correspond to practice, particularly with respect to the increasing use of lower drive currents to

minimize current density droop. Nevertheless, using a standard current (or power density, as measured in

Amps per centimeter squared, or A/cm2) at a fixed operating temperature and selecting devices within limited

ranges of CCT and CRI allows researchers to evaluate developments in emitter efficiency (including the

reduction of current density and thermal droop) and down-converter performance.8

Using these assumed operating conditions, in just 10 years, luminous efficacies have increased by a factor of

more than three, from less than 50 lm/W to approximately 165 lm/W. The principal reason has been

improvement in blue LED efficiency, although progress has also been made in phosphors (efficiency and

wavelength match to the human eye response) and package (optical scattering/absorption) efficiency. Despite

these improvements, there is significant remaining potential for improved efficacy. As illustrated by the

saturation values of the blue and yellow curves in Figure 5-3, luminous efficacies of approximately 255 lm/W

are believed to be practically possible for pc-LEDs.

For the color-mixed architectures an upper limit of 325 lm/W is considered achievable with technology

advancements discussed in this chapter. While the performance potential is high, today’s efficacies are much

lower than the pc-LED approach due to the inefficient green and amber direct emission LEDs. Panel (b) of

Figure 5-3 shows projections for power conversion efficiency of blue (440-460 nanometers, or nm), green

(530-550 nm), amber (570-590 nm), and near red (610-620 nm) direct emitting LEDs, again with a logistic fit

for projected performance, and with an upper limit of 90% power conversion efficiency.

In addition, Table 5-1 shows historical and projected LED package efficacy for warm white and cool white

phosphor-converted and color mixed LEDs.

8 For additional details regarding the specific operating conditions by which LED products are evaluated, see the notes described within Figure 5-3.

30

Figure 5-3. Efficacies and Efficiencies Over Time of White and Colored LED Packages

Note: All curves are logistic fits using various assumptions for long-term future performance and historical experimental

data. The data are from qualified products at the representative operating conditions of 25°C and 35 A/cm2 input current

density. They will differ from some commercial products, particularly those that operate at lower drive current densities to

minimize current droop.

• The upper panel (a) are the luminous efficacies of warm white (3000 K) and cool white (5700 K) phosphor-

converted LEDs and hypothetical color-mixed LEDs (CM-LEDs) with a CCT of 3000-4000 K. Luminous efficacies have

the typical units of photopic lumens of light (lm) created per input electrical Watt (We) of wall-plug power. Year 2017

commercial products reach approximately 180 lm/W for cool white PC-LEDs and approximately 160 lm/W for

warm white PC-LEDs. These values correspond to raw electrical-to-optical power-conversion efficiencies of

approximately 0.5 Wo/We.

• The lower panel (b) are the power-conversion efficiencies of direct-emitting LEDs at the various colors (blue, green,

amber, and near-red) necessary for CM-LED white light of highest source luminous efficacy and high color rendering

quality. Approximate future potential power-conversion efficiencies are depicted as a saturation at 90% for all

colors beginning in the years 2035–2040. The historical power conversion efficiencies of these sources were

combined and appropriately weighted to give the CM-LED LEDs and conversion efficiencies depicted in the upper

panel (a).

Table 5-1. Phosphor-Converted and Color Mixed LED Package Historical and Targeted Efficacy

Metric Type 2016 2017 2025 2035 Goal

LED Package Efficacy

(lm/W)

Cool White 160 172 241 249 250

Warm White 140 156 237 249 250

Color Mixed 90 100 196 288 330

Power-Conversion Efficiency ~0.5 Wo/We

31

5.2 Advanced Material Discovery and Engineering Exploration for LED Development

As discussed in the previous Section 5.1, the past 10 years have seen remarkable progress in LED luminous

efficacy, however LED technology still has significant room for improvement. While the pc-LED architecture

has seen the most rapid improvements, cm-LED architectures continue to be the architectures with the most

potential for efficacy. The BTO SSL Program goal for the white pc-LED is 250 lm/W, while for the white cm-

LED is 325 lm/W, which means there is a possibility for another 30% improvement after the phosphor

converted approach reaches its maximum efficacy.

However, progress towards improving the efficiency of cm-LED architectures requires improving efficiency in

the green-amber-red spectral regions, where there has been limited development. Possible solutions could arise

through use of current LED materials such as InGaN and AlInGaP, or through the development of new LED

materials. With this uncertainty in mind, two potential research directions are: advanced material discovery or,

advancements within the currently used LED materials systems.

Advanced material discovery is necessary to accelerate the rate at which new potentially viable materials are

detected. This process involves the evaluation of large numbers of material combinations through the use of

powerful computational techniques. Given the near incomprehensive volume of possible materials,

supercomputing enables researchers to apply techniques such as machine learning to problems in materials

discovery. Further R&D is needed to determine how advanced material discovery techniques could be tailored

to identify materials with potential for mitigating or by-passing the efficiency losses (both current density and

thermally driven) from the green-amber-red gap. This could potentially be done by defining and screening for

an irreducible set of properties, such as a material synthesis and electroluminescence efficiency within the

desired spectral region.

While advanced material discovery has the potential to identify suitable material compositions, engineering

exploration is then needed to test them in the laboratory and within the cm-LED architecture. For example,

several materials have currently been identified with potential for mitigating or by-passing the efficiency losses

from the green to red spectral regions – these include GaNP, BInGaN, ultrathin and/or ordered InGaN, and

perovskites. Engineering exploration of all identified materials would be significantly time intensive, therefore

efforts must be concentrated on the most promising materials systems. These select few need to be tested using

high caliber synthesis, materials and characterization techniques, linking results to SSL performance goals. In

addition, learnings from the engineering exploration process need to feed back into the advanced material

discovery to provide additional guidance that will improve the identification process.

These two research directions are inherently cyclical, with advanced material discovery translating into

insights into how best undergo engineering exploration. The engineering exploration then provides feedback

that can be used to better tailor the techniques and screening process for advanced material discovery.

Combined these methods have the potential to solve the green-amber-near-red gap among other SSL material

challenges.

5.3 Understanding Droop in LEDs

How “hard” an LED is driven has a very important implication for SSL – reducing cost. If luminous efficacy

(lm/W) can be maintained, the “harder” an LED is driven (in terms of input power per unit chip area), the more

lumens are output per unit chip area. Since chip costs scale approximately as chip area, more lumens are then

produced per unit chip cost. This has historically been the principal motivation for eliminating or reducing blue

LED efficiency droop: the maintenance of high luminous efficacy as LEDs are driven harder, which enables

more lumens per LED cost.

However, as chip costs have decreased they have also become a smaller fraction of overall LED package cost.

Thus, the magnitude of cost impact of eliminating droop has declined. Nonetheless, eliminating efficiency

droop still offers significant benefit, particularly when accounting for overall LED system-level costs, which

can be reduced if boards, heat sinks, optics, etc., can be made physically smaller per lumen of light output.

32

5.3.1 Blue LEDs

The efficiency of blue LEDs has improved enormously over the past decade. Leading research has

demonstrated blue LEDs that exceed 80% external quantum efficiency (EQE), but only at relatively low

current densities. LED efficiency is still limited at high current density due to a phenomenon known as

efficiency droop or current density droop. Operation at higher current densities is desirable to maximize the

light emitted from the chip, thereby lowering the cost per lumen of LED lighting products.

There are different physical mechanisms that impact efficiency at different current densities, as indicated in

Figure 5-4. At low current densities, the number of defects in the material has a significant impact on

efficiency, where Shockley-Read-Hall (SRH) nonradiative recombination dominates. At higher current density

operation, Auger recombination dominates, which is a non-radiative carrier recombination process which

increases nonlinearly with carrier density and hence current density. Possible approaches to circumvent Auger

recombination losses include increasing the rate of competing radiative recombination (either through

composition/geometry engineering or through use of alternative recombination mechanisms such as stimulated

emission in laser diodes) or decreasing carrier densities in the active region (either through band-

structure/transport engineering or through alternative geometries such as stacked active regions connected via

tunnel junctions). The key to any of these approaches is to understand and control the complex epitaxial

materials synthesis process in order to maintain the material quality within the LED structure [12].

The amount of Auger recombination is controlled by the Auger constant in each quantum well (QW) of the

LED active region and the carrier density in each QW, so it is important to have uniform current injection into

each QW. The LED epitaxial design can be changed to increase the carrier transport to get uniform injection

into each quantum well, as illustrated in Figure 5-4. The problem is that the improved heterostructure leading

to uniform carrier injection into the active region, leads to growth conditions that increase the SRH

nonradiative recombination. While progress has been made in this area through funded R&D projects, further

research in InGaN epitaxial growth is required to continue balance the material quality with the improved

heterostructure design for carrier transport [13].

33

Figure 5-4. Blue LED EQE vs. Current Density and Schematic of LED Quantum Well Valence Band [14]

Note: Blue LED external quantum efficiency as a function of current density (left); the shaded regions indicate the

dominant recombination modes. Schematic of the quantum well valence band of an LED showing carriers piling up in the

p-side QW (top right) and showing uniform hole injection (bottom left).

5.3.2 Green LEDs

Although the InGaN alloy can theoretically cover the whole visible spectrum, its quantum efficiency drops

rapidly above 500 nm as emission shifts from blue to green. Considering the long wavelength side of the

visible spectrum, the AlGaInP materials system can provide high-performance red LEDs though the efficiency

drops steeply in the amber region [15]. This phenomenon is known as the ‘green gap’ and is illustrated in

Figure 5-5. The low efficiency of green LED is particularly critical, since ultra-efficient white LEDs based on

color mixing require a green LED emitter with wavelength around 540 nm — right near the center of the

‘green gap.’

Figure 5-5. Spectral Power Densities of State-of-the-Art Commercial LEDs vs. Wavelength

Notes: Dashed lines are guides to the eye, illustrating the “green gap”: the decrease in efficiency from the blue to the

green-yellow and from the red to the green-yellow. Data is for operation at 85°C and has been “stylized” into Gaussian

spectral distributions using efficiencies, center wavelengths and spectral linewidths from the Lumileds Luxeon C Color

Line Datasheet DS144 (2018 02 19).

The source of the efficiency drop in the AlGaInP materials system is due to the transition from a direct bandgap to an indirect bandgap in the amber/green spectral region. For InGaN, the materials are less efficient

in the green due to the combined effects of high indium compositions (materials challenges), polarization

fields (less electron hole wavefunction overlap), and greater Auger recombination. The current density droop

QW

++ ++

VB

QW

++++

VB

34

problem for green LEDs is even more severe than for blue LEDs. Figure 5-6 shows a schematic of the carrier

distribution in a blue LED active region and a green LED active region. The carrier distribution in the green

LED active region is poor due to larger energy barriers slowing vertical transport in active region. The

increased barriers to carrier transport also result in lower electrical efficiency as compared to blue LEDs due to

higher forward voltage relative to its photon energy [14].

Figure 5-6. Schematic of the LED Quantum Well Valence Band [14]

Notes: Schematic of the quantum well valence band of an LED showing carrier distribution in today’s state-of-the-art blue

LED QW active region (top) and the carriers piling up in the p-side QW for a green LED active region (top right) and

showing uniform hole injection (bottom).

To address the current density droop in green LEDs, more R&D on improving carrier transport between QWs

is critical, even more so in green than blue LEDs. However, the biggest challenge is that most LED

heterostructure changes that improve carrier transport hurt the material quality — again this is exacerbated for

green LEDs relative to blue. Fundamental research in droop mitigation strategies should benefit both blue and

green LEDs, though the challenges are magnified in the green spectral region.

5.3.3 Thermal Droop

Thermal droop in LEDs is simply the reduction of the optical power when the temperature is increased, which

limits the efficiency of LEDs beyond that attributed to current density droop. Thermal droop is important in

commercial devices since the temperature increases at the typical operating conditions in LED luminaires.

Some commercial white LEDs are rated for operating up to 150°C, though devices running at 150°C can lose

up to 25% of optical power, compared with room-temperature operation. The light output decline is more

severe for the AlGaInP materials system where the optical power can drop 70% at 150°C. Figure 5-7 shows

some typically thermal droop behavior for various color LEDs.

VB

QW

++++ +

+

p-sideActive region with QWs

emitting at Blue WL

Active region with QWs

emitting at Green WL

Active region with QWs

emitting at Green WL

QW

++++

VB

35

Figure 5-7. LED Efficiency as Function of Junction Temperature [16]

Note: LED efficiency declines as junction temperature increases. AlGaInP and AlGaAs LEDs experience the greatest drop.

Thermal droop occurs because of temperature-dependent semiconductor properties that cause non-radiative

recombination and carrier loss. Researchers have been looking for the origin of thermal droop in InGaN LEDs.

Work done by researchers at the University of California Santa Barbara show that when blue LEDs are

operated at elevated temperatures, they demonstrate an increase in electrons lost via carrier leakage and/or

overshoot. This increase of leakage and/or overshoot coincides with the onset of the decrease in light output at

~75°C, a temperature range at which LEDs are commonly operated. These results are consistent with the

expected onset of the thermal droop that has been widely reported in scientific literature [17]. New InGaN

LED heterostructure designs are needed that can minimize the carrier overshoot at elevated temperatures while

maintaining the materials quality and high efficiency.

Thermal droop in AlGaInP LEDs is much greater than in InGaN LEDs. This is due to the materials properties

in the semiconductor system. AlGaInP has small band offsets which can lead to significant carrier overflow

with increasing temperature, especially for the shorter wavelengths such as amber. Research into new strain

engineering approaches for epitaxial growth of the active region is a promising approach for improving the

carrier confinement and reducing carrier overflow.

5.4 Advanced LED Architectures

Advanced LED device architectures have the ability to improve efficiency or improve the device’s operating

ranges. These can lead to improvements in current density droop or provide desirable device performance,

such as high luminance, that is not achieved with conventional LEDs.

5.4.1 Droop Mitigation

There are several approaches to reducing or mitigating the impact of droop. One approach is to redesign LED

active regions to minimize carrier density within them, as discussed in the previous Section 5.3. This reduces

droop; however manufacturers have discovered that it is very difficult to maintain LED material quality with

these low-droop designs.

There are also device architecture approaches to mitigating droop – such as using a laser diode (LD) to

mitigate droop. In LDs, droop is eliminated when lasing occurs since all excess carriers are consumed by

stimulated emission, thus reducing the availability of carriers for the non-radiative Auger recombination

processes. This can allow for high flux density and higher wall-plug efficiencies than LEDs at very high

36

current density operation. LDs have clamped charge carrier density, so droop does not exponentially increase

at higher operating current, however, with lasers there is also a tradeoff between peak efficiency and droop

reduction. Researchers are working on both the peak efficiency of lasers and ways to integrate them into

practical lighting products.

As seen in Figure 5-8, an interesting insight involved the so-called “valley of droop” – the region of current

density which is high enough that significant LED droop occurs, but low enough that laser diodes do not yet

lase. Up until recently, it was thought that current densities associated with the valley of droop were optimal: if

LEDs could be driven that “hard” while circumventing droop their photons would be less expensive; if lasers

could be driven that “soft” while still lasing resistive losses would be lower and their efficiencies higher.

While the current densities associated with the valley of droop would still be desirable, two trends make it

economical to consider on both sides of the valley of droop. First, because the cost of the chip, particularly the

cost of the epitaxy, continues to decrease, larger chips driven at lower current density might soon be more

economical. Thus, it is of interest to continue to increase peak efficiencies for low current density operation.

Figure 5-8. PCE Vs. Current Density for a State-of-The-Art LED and LD Emitting at Violet Wavelengths [18]

Notes: Plot of power conversion efficiency (PCE) vs. current density for a state-of-the-art LED and LD emitting at violet

wavelengths highlighting the ‘valley of droop’ cross-over between source types.

Second, directional light is becoming increasingly important because it improves photon utilization efficiency.

There is a premium placed on small, low etendue sources that can be spatially focused and directed. This is the

province of high current densities: blue laser diodes beyond the valley of droop, and blue LEDs driven as far

into the valley of droop as possible. Further R&D for laser lighting includes increasing the wall plug efficiency

(WPE) of the LD from the current 30-40% range to 60% (LED level).

Finally, new architectures are being explored that could enable the effective straddling of the valley of droop

simultaneously in a single structure: stacked tunnel-junction (TJ) series connected LEDs. Essentially, it would

create multiple LEDs in series, which would increase voltage while keeping current low. This would enable

higher light output from an area of LED material, while keeping the applied current — and resulting droop —

low, as illustrated in Figure 5-9.

37

Figure 5-9. Schematic Band Diagram of a Stacked Active Region LED with Tunnel Junctions (left); the External Quantum

Efficiency at the Peak LED Operating Current (right) [19]

While research into TJs has increased in recent years, several challenges remain. The increased voltage drop

that results from the increased stack voltage can be an issue in TJs and needs to be reduced. Additionally, there

are issues associated with activating the p-type dopant in buried active regions grown by metal-organic

chemical vapor deposition (MOCVD) and absorption when using InGaN TJs. Moreover, developing growth

processes for growing high-quality TJs is required to keep defect densities low and minimize negative impacts

of subsequent LED junctions. Alternatively, growth processes such as molecular beam epitaxy (MBE) can be

used to overcome some of the growth and activation challenges facing MOCVD (though the added cost of a

second growth technique must be overcome).

5.4.2 High Luminance

While improving the efficiency of emitted light from an LED has been a big focus in the LED industry, how

that light is delivered to the lighting application is equally as important. Some lighting applications, such as

spot lighting, require a very narrow beam of light to illuminate the desired object. If the light is not focused in

a tight beam, a significant amount of light generated from the source is not useful, thus lowering the optical

delivery efficiency of the luminaire system.

The directionality of the light source also plays a big role in the efficacy of a light source. The ‘harder’ you can

drive a light source, the more light you can generate out of a given area, thus increasing the luminance

emittance. Luminance emittance is the luminous flux per unit area emitted from a surface expressed in units of

lm/mm2. When you can increase the lumen emittance, optical source size for a given lumen output can

decrease. The smaller the optical source size, the smaller the illuminated area can be for a given size of

package/luminaire optics. Equivalently, the smaller the package/luminaire optics can be for a given size of the

illuminated area. Thus, in directional illumination, where the spatial profile of the illumination area is tailored,

driving LEDs harder to achieve a smaller source size becomes more important.

However, just as efficiency droop causes the trade-off between cost and performance, as discussed in Section

5.3, it also causes a trade-off between luminous efficacy and luminous emittance. Figure 5-10 compares

several representative state-of-the-art 2017 commercial white light packages and shows the wide span in

efficacy and luminous emittance. As input current density increases, luminous efficacy decreases while

luminous emittances increase. At the extreme top left is a mid-power white LED package driven at 0.7 A/cm2

(shown in dark green text), while at the middle right is a high-power white LED package driven at 35 A/cm2

(blue text). Also shown at the extreme bottom right is an estimated point for a laser diode (LD) white light

package (bright green text).

38

Figure 5-10. Luminous Efficacy vs. Luminous Emittances for State-of-the-Art Commercial White LEDs

Note: The dashed line is an empirical fit using the Equation given in the text.

A log-linear fit to the data points gives the empirical equation:

η = 225 lm

𝑊 – 46.4 x log10 [

𝑀𝑉

1 𝑙𝑚

𝑚𝑚2

]

This above shown equation can be thought of as defining the current trade-off between luminous efficacy, η,

and luminous emittance, MV. Additional research is needed that focuses on materials and device architectures

that go beyond the current state-of-the-art to enable both high luminous efficacy and luminous emittance – as

demonstrated by the upper right quadrant in Figure 5-10. Research areas include further reductions in LED

efficiency droop, down-converter materials improvement to provide high efficiency and stability at higher

luminance, packaging materials improvement to prevent degradation at higher optical flux densities and

temperatures, and optical design for angular uniformity of color.

5.5 Optical Down-Converters

State-of-the-art LED lamps and luminaires are predominantly based on phosphor-converted LEDs (pc-LEDs).

The phosphors used in these pc-LEDs result in an emission with broad linewidths, which in turn limits their

overall spectral efficiency or luminous efficacy of radiation (LER). The broad linewidth is particularly

significant for the red spectral region since the broad emission results in a larger portion of the overall light

39

distribution to be emitted in regions of the visible spectrum where the human eye is less sensitive. This portion

becomes larger as the CRI increases, because a higher CRI puts more stringent demands on the amount of light

emitted in the red wavelength range at the edge of the visible spectrum. However, because pc-LEDs emit a

larger portion of their light in those regions, lamps or luminaires made with 90 CRI pc-LEDs have lower

efficacy than those made with 80 CRI pc-LEDs due to this spectral inefficiency. This efficacy gap must be

minimized to stimulate greater adoption of 90 CRI, pc-LEDs for lighting.

5.5.1 Narrow-Band Phosphors

Typical nitride or oxynitride red LED phosphors have a wide emission linewidth near 100 nm full width at half

maximum (FWHM). This causes a significant spillover of light into the deeper red wavelength range, where

the human eye is less sensitive, and is a significant contributor to the spectral inefficiency of current pc-LED

white light. Figure 5-11 illustrates this behavior by comparing a white LED using a 110 nm FWHM broadband

red phosphor with a CCT of 3000K, a CRI ≥90, and an R9>50 to a white LED (with similar color qualities)

using a red phosphor with bandwidth of 30 nm. A 22% improvement in spectral efficiency is gained by

replacing the red broadband phosphor, which reduces the wasted emission in the deep red and infrared (IR)

wavelength ranges (beyond 650 nm) [20].

Figure 5-11. Spectrum Comparison of 90 CRI PC-LED, 90 CRI Conventional Red Phosphor LED, and Human Eye Response

[20]

There have been recent developments in the field of narrow red down-converters. GE continues to release

lighting products that feature its narrow red phosphor, “KSF,” under their “Tri-Gain” brand [21].9 These lights

exhibit excellent color quality and high efficacy due to the narrow red emission spectrum of the phosphor.

While this phosphor was demonstrated several years back, materials refinements have continually improved its

long term behavior. Such improvements include a smaller color shift in LED packages and stronger lumen

maintenance stability under high blue flux densities, as seen in Figure 5-12 [21]. Similarly, Lumileds has

commercialized mid-power LED packages that use its “SLA” phosphor to provide narrow red emission and

enable good color quality and high efficacy [20].10

9 KSF, or K2SiF6:Mn4+, is a potassium fluorosilicate phosphor. 10 SLA, or Sr[LiAl3N4]:Eu2+, is a nitridoaluminate compound.

40

Figure 5-12. Light Loss of Phosphors in LED Packages under High Blue Flux Densities (left) and Color Shift under Stressed

Operating Conditions (right) [21]

While significant improvements have been made to narrow-band red phosphors over the past several years,

opportunities still exist to improve material synthesis and composition to result in fewer materials defects and

to allow for higher activator manganese (Mn) concentrations, which can reduce the amount of phosphor

materials needed on the LED. These additional improvements would lead to lower phosphor volumes at the

same color point currently in a comparable LED. Further reliability improvements are also desirable to run at

higher operating fluxes and temperatures.

5.5.2 Quantum Dot Down-Converters

Quantum dots (QDs) have long been targeted for use as down-converters in LEDs due to their combination of

two unique emission characteristics: tunability of wavelength and narrow emission linewidths. These quantum-

confined semiconducting nanocrystals are made of inorganic semiconductor material and commonly “grown”

using colloidal synthetic chemistry, with electron and hole confinement, that results in unique optical

properties. Colloidal QDs feature a tunable band gap that can span the entire visible spectrum with nanometer

scale resolution by adjusting the particle size and a narrow FWHM owing to the direct transition from the band

gap edge. Until now, QDs have not gained much traction as a drop-in solution into the LED package because

the LED operating temperature and blue flux intensities result in strong thermal quenching and fast photo-

degradation. R&D progress in this area has been made, though, with Lumileds’ commercialization of a mid-

power LED package using red QD down-converters (combined with phosphors) this year [22] [15].

As with narrow-band phosphors described in Section 5.5.1, the use of QDs as down-converters can provide

improved spectral efficiency gains by reducing the wasted light emission in the deep red and IR portions of the

spectrum. Red QDs used in combination with a conventional phosphor material can improve LED conversion

efficiency by 5% to 15% over commercial pc-LEDs between CCTs of 2700 Kelvin (K) to 5000 K [23].

Lumileds LEDs with the on-chip application of QDs can operate where the QD temperature exceeds 100°C

and the blue flux intensity reaches 0.2 W/mm2 in mid-power packages. These achievements in QDs

demonstrate the essential reliability requirements for use in commercial applications [24].

However, the current high-performance QDs commercialized in LEDs contain a small amount of cadmium

(Cd). The use of Cd in electronic devices is regulated by the European Union (EU) under the Restriction of

Hazardous Substances (RoHS) Directive; Cd use is limited to 100 parts per million (ppm) in the smallest

homogeneous component of an electronic device containing the metal. For on-chip LED usage, the smallest

homogeneous component is the down-conversion layer consisting of the QDs, other phosphors, and the silicone binder that is deposited inside the LED package. The exact concentration of Cd depends on multiple

41

factors, such as the LED package design and the final color point, but it has been estimated to range between

150 and 500 ppm [25].

While Cd-containing QDs provide the best performance to date, there is still the need to evaluate alternative

Cd-free QDs due to the regulatory requirements on Cd use. The most advanced Cd-free QD technology is

currently InP-based QDs; however, currently the FWHM of the emission and environmental stability is not to

the level of their Cd-containing counterparts. The FWHM has improved the past few years and is now

approximately 40 nm for green and 50 nm for red, nearing the BTO target of 30 nm FWHM [26]. The progress

in the last few years has come from better materials design, but stability is still a large hurdle that requires

further research and development. Other potential QD systems (that do not contain Cd and will be ROHS

compliant) are perovskites and CuSeS QDs, which are still in the early stages of development and require more

work to assess the performance levels and stability.

Beyond creating QDs with the required performance properties and reliability behavior for incorporation in

LED packages, the ability to manufacture large-scale batches of QD material is critical for use in SSL. One

significant hurdle in QD synthesis is controlling the size of the actual QDs. Slight diameter changes will result

in wavelength changes in the down-converter, as illustrated in Figure 5-13. When the ensemble of QDs with

slightly varying diameters is applied in an LED package, the emission FWHM can broaden. New synthesis

techniques can help improve the layer-by-layer synthesis, which is difficult to consistently control.

One effort to potentially significantly improve the scalable synthesis of high-performance QDs employs a

convergent (rather than linear) approach that uses a single-step heterostructure synthesis. This creates graded

alloy QD architectures using tunable reaction kinetics of a set of precursors. Reliably dictating QD size,

concentration, and monodispersity requires well-controlled precursor conversion. Research is underway to

prove out the synthesis reproducibility, QD performance, and reliability using new colloidal synthesis.

Figure 5-13. Emission Wavelength of CdSe QDs as a Function of Dot Diameter [26]

42

QDs for on-chip LED application have made remarkable advancements over the past few years. CdSe-based

QDs have been released in an initial LED package product this year, demonstrating their commercial viability

[22]. While the progress has been promising, more research and development work is required to advance

understanding in high-efficiency, on-chip QD down-converters to match or exceed performance of

conventional on-chip phosphor materials. In addition, further development of QDs that do not contain heavy

metals (such as Cd or Pb) or scarce materials is needed for the changing regulatory requirements on these

materials.

5.6 Additive Fabrication Technologies for Lighting

Over the past few years, additive manufacturing has been a growing area of interest for SSL product

prototyping and manufacturing. Additive manufacturing is a fabrication process where a 3D object is created

by computer-controlled deposition of material (in a layer by layer approach) based on a computer-aided design

(CAD) model. 3D printing is one common example of additive manufacturing. It can be more efficient than

traditional “subtractive” manufacturing approaches, such as milling, grinding, and polishing, which involve

removing material to achieve the desired form, either for the product directly or for making molds and tooling.

Some of the key benefits of additive manufacturing include:

• Complex shapes can be made that are not possible with traditional manufacturing;

• Opportunities to use mixed material types in the same package;

• Lower energy intensity by eliminating production steps, using substantially less material, and producing

lighter products;

• Minimized initial equipment investment cost – i.e., no tooling is required, offering flexibility of shapes

and reduced inventory with the same equipment;

• Easier to iterate product variations or functional form-and-fit processes and testing; and

• On-demand manufacturing capabilities for projects with shorter lead times.

Additive manufacturing applies to many different aspects of the SSL supply chain and manufacturing

processes. Some of these areas include:

• Wafer scale packaging, including down-converter and encapsulant deposition;

• Power supply component and module manufacturing;

• Rapid creation of tooling for optics, heat sink, or housing manufacturing; and

• Flexible production of optics or lighting fixtures.

The primary use of additive manufacturing in SSL has been for rapid prototyping on new product design

concepts. 3D printing enables the design of custom fixtures with improved visual appeal, better functionality,

and reduced fixture costs. The Philips Lighting (now Signify11) Telecaster Program has been developing

additive manufacturing for numerous styles of light fixtures, including 3D-printed decorative pendants, track

spots, downlights, and large high bay fixtures. Figure 5-14 shows three examples of 3D-printed light fixtures.

11 As of May 16, 2018, Philips Lighting is now referred to as Signify. Additional information on the announcement of the name change can be found at:

https://www.signify.com/en-us/about/news/press-releases/2018/20180516-philips-lighting-is-now-signify.

https://www.signify.com/en-us/about/news/press-releases/2018/20180516-philips-lighting-is-now-signify

43

Figure 5-14. Images of 3D-Printed Lighting Fixtures [27]

Beyond the use of additive manufacturing to make luminaire housings, this technique has been used to create

the functional components of luminaires, such as optics. These optical structures are made from a UV-curable

polymer ink and cured by UV lamps in the print head upon each pass of printed droplets, as illustrated in

Figure 5-15. This method allows geometric and free form shapes to provide the desired optical control

features, while it simultaneously eliminates the expense of molds and tooling and enables just-in-time

manufacturing.

Figure 5-15. Deposition of Droplets by UV Print Head onto Substrate Material (left-top). Droplets of Polymer Are Allowed to

“Flow” under Surface Tension before Curing with UV Light, giving Smooth Surfaces Needed for Optics (left-bottom). Array

of Micro-Optic Lenses (right). [28]

One of the biggest challenges with implementing 3D printing further into the SSL value chain is the

development of printable materials with properties specific to lighting applications – optical, electronic, and

thermal properties. In this space the manufacturing advancements achieved by EERE’s Advanced

Manufacturing Office can be leveraged; however, application-specific manufacturing R&D is necessary. For

example, there are challenges achieving the appropriate thermal conductivity of a heat sink using a polymer-

based ink with conductive fillers. While these materials can be used to print a heatsink, the thermal

conductivity falls short of the performance seen with aluminum heatsinks. Studies have been carried out

printing electrical traces for printed circuit boards (PCBs), and while they can be printed, the resistivity of the

traces are higher than copper [29]. While proof-of-concept demonstrations exist for the use of additive

manufacturing in many areas of the SSL value chain, more research and development is required to develop

Pendant Downlight Highbay

44

printable materials with the sufficient properties to replace existing manufacturing approaches in electrical,

thermal, and optical components.

Another area of interest for additive manufacturing in the SSL value chain is to create tooling using 3D

printing. The lead time for tooling for molding or stamping processes often takes 10-12 weeks to be created.

3D printing has the potential to reduce the lead time significantly and create tooling in 2-4 weeks. This allows

for a shorter product development cycle and quicker pilot line development, and the concept has been used to

prove out the 3D printing of cars. The use of additive manufacturing in creating tooling has the potential to

create efficiency gains with SSL product manufacturing.

5.7 Advanced LED Lighting Concepts

Developing lighting system architectures that take advantage of the unique properties of LEDs can maximize

the vast potential of LED lighting over traditional lighting sources. LED lighting provides features such as

high energy efficiency to color quality, compact and unique shapes, and other non-energy benefits (e.g.,

physiological responses to light). Improved lighting application efficiency (discussed in Section 4.1) can help

further luminaire performance levels and provide advanced lighting values (e.g., human physiological benefits

as demonstrated by spectrum and intensity levels appropriate for engaging these responses). Concepts that

demonstrate improvements to lighting application efficiency should be a focus of future research and

development.

Color-tunable LED luminaires provide the ability for spectral and intensity tuning, which realizes some value-

added features available from LED lighting for applications such as hospital lighting to horticulture. These

color-tunable lighting systems, also referred to as RGB, RGBA, spectrally-tunable, or color-changing, usually

have three or more different LED primaries that can be individually varied in light output power to create a

mixture of light that is white or a saturated hue. The individual LEDs used in a full-color-tuning mixture can

be direct emission narrow-band LEDs (producing a narrow range of blue or red, for example), or they can be

monochromatic with phosphor coatings that produce a slightly wider spread of color (e.g., a “mint” green LED

is a phosphor-coated blue). Usually the different monochromatic LED colors include red, green, and blue (i.e.,

RGB, the primary colors of light), but these can be augmented with amber (A) or other monochromatic colors.

The minimum number of LED colors is three for full-color tuning, though four-, five-, and seven-color

systems are also on the architectural lighting market, and some sophisticated color systems use additional

unique colors of individual LEDs.

One unique advantage of this type of color-tuning is the ability to move the color point off the blackbody locus

or, in other words, to move beyond different CCTs of white light toward light with a distinct color. For

example, such a product could provide 4000 K light in an office during the day and then tune to a different

color for a purple-themed party in the evening. This makes full-color-tunable products well-suited for theater,

theme park, and restaurant applications. Another advantage of full-color tuning is the ability to match the

chromaticity of any other light source. Controlling the colors of individual LEDs introduces the option of

tuning the spectrum to enhance colors for retail applications – for example, a lighting display can be tuned to

make a floral arrangement appear rich and full in color.

The four-color RYGB cm-LED architecture, in which all colors are generated by direct LEDs, can

significantly improve efficacy since it will remove the fundamental Stokes losses associated with down-

conversion by phosphors or QDs. As indicated by the dashed grey line in Figure 5-3, ultimate upper potential

of a four-color RYGB cm-LED can be on the order of 325 lm/W, limited only by the anticipated 80% to 90%

efficiencies of the actual LEDs and for the losses when mixing the pure source colors to create white light. At

present, the low external quantum efficiency of green/amber LEDs lower the overall system efficacy of a

color-mixed system compared to a white pc-LED system. As discussed earlier, the green and amber LED

efficiency must be increased to achieve the combined potential of maximized high efficacy and full color

control.

With the addition of color-tunable luminaires to the market, color mixing has become a growing challenge.

New color mixing optical schemes (e.g., refractive + lighting guiding optics) are needed in luminaires to help

45

provide low-profile optics and efficient mixing optics and to reduce the cost compared to the traditional

volume mixing chambers in today’s luminaires. The development of efficient optical elements should be

considered to improve the efficacy of color mixing.

In addition to more complexity in color mixing, more complexity is required in the control strategies for fully

color tunable LED luminaires. This wide variability of full-color-tuning requires a user interface that is more

complicated than a simple slide dimmer. These tunable luminaires usually require full, undimmed power to be

delivered to the LED drivers in the luminaire, typically delivered using standard building voltage (such as

120V, 277V) using conventional hard-wiring techniques. They generally require separate instructions for the

intensity of each color of LED in the optical mix (such as warm white, cool white, red, green, blue, amber, etc.,

depending on the product design), with instructions for dimmed level and color sent over separate wires or

through wireless signals from the user interface to the driver, or from the user interface to the black-box device

that processes control signals and converts them to driver instructions. A control protocol such as DMX or

DALI can address that the luminaire must be powered separately from the intensity and color control signals.

DMX was originally created for the theatre industry, but today DMX is also widely used for dynamic lighting

since it is fast – changes in intensity and color can be made virtually instantly.

While these color control strategies are available today, there is a need to reduce the complexity with new

driver architectures. Reducing the LED drive output channels from individual to combined control can be one

such improvement. Another area where there is room for improvement is how to implement the color mixing

logic. Moving the color mixing science and the LED output control to an on-board LED channel management

has promise to reduce this complexity [30]. More research and development are needed to create very efficient

compact multichannel drivers for color tunable systems.

5.8 LED Power and Functional Electronics

The LED driver is a critical component to the LED luminaire since it powers the LEDs. The driver accepts

input power of various types, including conventional alternating current (AC) line power, as well as direct

current (DC) power from DC micro-grids or Power over Ethernet (PoE). From there the driver outputs

voltages and currents compatible with the LED packages, over single or multiple channels, and may

incorporate control functions such as dimmability and color-temperature tuning. The two key aspects of the

driver are its reliability and performance, where performance can include efficiency, flicker, surge rating,

enhanced lighting functionality, non-lighting multi-functionality, as well as size, weight, and power level

(SWaP).

5.8.1 Driver Performance

The key performance metrics of drivers focus on their ability to transform power appropriately and efficiently,

while protecting downstream components from power surges and poor incoming power quality. These

performance metrics for LED drivers include efficiency (both full power and dimmed), dimming level,

absence of flicker, surge protection, size, weight, accommodation of multiple channels and alternative input

power.

On/off/dim capability is important as lighting becomes connected and adaptive to user needs and preferences.

These functions need to be performed at high driver efficiencies, a challenge in today’s drivers where

efficiency drops in the dimmed state. Absence of flicker is important for any light source but can be

challenging due to a lack of standard definitions for basic quantities such as percent flicker and flicker index.

This is further complicated in part because of new types of flicker such as CCT flicker in color tunable lighting

systems. Accommodating multiple channels is important for color tuning and/or driving multiple LEDs and

LED strings. The ability to utilize alternative input power includes inputs such as DC micro-grids or PoE,

which will prove vital for multifunctionality. PoE is a fast-evolving area, as IEEE PoE standards are updated to

enable lighting applications by providing higher maximum power per port and per device.

Another overarching feature is the size, weight and power of the driver. In virtually all use cases, a compact

driver form factor is better; however, in some use cases it is essential to the functionality of the luminaire. In

46

general, making luminaires smaller would enable greater flexibility and density of luminaire placement, which

in turn would enable lighting architects to more freely control lighting scenes and provide denser spatial

coverage of sensors. Thus, an important challenge to be addressed is continuing improvement in SWaP, even

while sustaining the performance metrics outlined above. A big challenge is maintaining high efficiency and

small, light drivers over a large operating power range. Integration of wide-bandgap semiconductors

components into the driver have the potential to address a number of these performance metrics and is a

potential R&D path. Gallium nitride (GaN) or silicon carbide (SiC) wide-bandgap semiconductors with higher

breakdown voltages and greater robustness against power surges may enable two-stage drivers to be reduced to

one stage. Furthermore, wide-bandgap semiconductors enable higher switching speeds for voltage

transformation. All of these benefits have the potential for size reduction and efficiency improvements.

Although SiC is currently ahead, both SiC and GaN are much less mature than Si, so costs are relatively high.

Further research is needed to develop consistency, improve reliability, and to reduce cost of SiC and GaN-

based components. One of the advantages of GaN power electronics is that it is able to draw on the

considerable existing knowledge and manufacturing base established by InGaN/GaN-based LED lighting.

Because of this, it would be coming full circle for LED lighting to in turn benefit from the incorporation of

GaN into LED drivers. Indeed, because they share the same materials platform, a long-term opportunity could

be integration of GaN power electronics with InGaN/GaN LEDs. Such monolithic integration brings

challenges: potential incompatibilities in some of their epitaxial growth and fabrication processes, as well as an

inability to bin and match electronic and optoelectronic characteristics after separate fabrication. But such

monumental integration also brings opportunities: the pixelated light source discussed above might be most

elegantly realized with GaN-based display drivers integrated underneath pixelated LED light sources.

5.8.2 Reliability

Typically, the driver is the first component of a luminaire to fail. By and large, this is because LEDs are so

intrinsically reliable that drivers are the resulting weakest link. Driver reliability in some cases is not even as

robust as it was for earlier generations of traditional lighting, such as a copper-wound ballast system used for

high-intensity discharge (HID) lighting in industrial spaces. This is because power surges and other electrical

events that cause abnormalities in power quality can damage LED lighting components more so than

traditional lighting systems. While this is not a problem unique to lighting, as more fragile components are

introduced into the SSL system, protecting LED luminaires from poor power quality becomes more important.

Current surge protection systems are built around larger events, meaning that several smaller events or

transitions can get though surge protection systems and these load transitions cause field failures when the

power quality is poor.

Driver reliability is an area that presents a significant opportunity for improvement, including fundamental

reliability limitations of many of the subcomponents of the driver, such as electrolytic and film capacitors, and

to do so in a manner consistent with the ongoing trend to higher performance discussed in the previous Section

5.8.1. Another goal would be to develop a greater degree of power conditioning, especially as fragile

components are introduced into the SSL system due to the need for improved performance, particularly those

involved in multi-functionality. Currently, most current surge protection systems are designed to block larger

events, but not smaller events, which can accumulate over time and eventually cause damage to downstream

components.

A closely related challenge is to develop predictive driver reliability models and metrics. Current metrics, such

as mean time between failures (MBTF) for individual components, are considered inadequate. Therefore,

developing additional metrics to define failure, and ways to predict them would be beneficial to the SSL

industry. Metrics to describe performance features such as driver efficiency, maximum temperature rise over

ambient and how these change over time are also desirable. Coupled with such models and metrics would be

standard highly accelerated reliability testing protocols that can return results quickly, within a matter of

weeks.

47

Further research is needed to improve driver temperature performance, surge rating, reliability and cost. Solid-

state component integration into the driver should be explored as a more robust alternative since solid-state

drivers can simplify the part count and reduce failures. It would also improve the surge rating and reduce the

driver size. Moving GaN or SiC-based power electronics has the potential to improve the efficiency and

reliability, though today these solid-state components are still very costly and further research is required in the

electronics industry to improve the defect count and reduce cost. Establishing the reliability for GaN and SiC

components and the impact on driver reliability is an important opportunity.

5.8.3 Enhanced Functionality of Drivers

Enhanced lighting functionality will be a vital driver feature for future deployment of connected lighting with

advanced capabilities and will enable programmable control of that functionality. Real-time control of light

placement is an important such enhanced functionality. For example, optical beam shaping through digitally

controllable liquid-crystal lenses could enable significant improvement in the use efficiency of light, by

tailoring, in real-time, the lighting field of view to the user field of view. In another example, pixilated beams

could enable not only similar improvements in use efficiency of light, but also enable augmented reality that

highlights salient features of a user’s environment or provides other information to the user. Taken to its

logical limit, augmented reality would be a form of illumination and display convergence, which would require

drivers with video-display-like driver capability.

Finally, with the advent of connected lighting, lighting fixtures may well become the most ubiquitous grid-

connected end-point in the Internet of Things, with opportunity for many desirable new functionalities to be

embedded into the fixture. In the short term, separate drivers may be used for these new functionalities.

However, in the long term, there may be opportunity for integrated drivers that drive both the LED as well as

these new components. One new functionality is communication via Li-Fi, with its need for high-speed

modulation, interoperability, and end-to-end security requirements. Another new potential functionality is

sensors for monitoring all aspects of the environment including sound, light, temperature, chemicals, motion,

human presence, perhaps even LIDAR-based 3D mapping. The complexity of these offerings becomes

enormous as each has its own requirements for interoperability and end-to-end security.

48

Directions in OLED Science OLED technology is steadily improving with commercial products now available that reach performance goals

of high efficacy, long lumen maintenance lifetime, and good color quality. OLED lighting configured in a

commercial office setting is shown in Figure 6-1.

Figure 6-1. Acuity OLED Luminaires in Office Space

Source: PNNL, “OLED Lighting Products: Capabilities, Challenges, Potential,” May 2016 [31] Photos courtesy of Aurora

Lighting Design

Further, bendable panels have been commercialized. US-based OLED panel manufacturer, OLEDWorks has

made great strides over the past year with the announcement of their Brite 3 panel family. At standard

luminance of 3,000 cd/m2, the warm white (3000K) panels demonstrate 85 lm/W and L70 of 100,000 hours.

This represents an impressive 35% increase in efficacy over previous generation Brite 2 panels which

demonstrate 63 lm/W, along with a substantial doubling of lifetime (L70). Color quality of panels remains

impressive for the Brite 3 products, which offer CRI >90 and R9 >50. Progress towards flexible panels has also

been made with the introduction of a uniaxially bendable panel, the Brite 3 Curve, also called the BendOLED.

This product is made using Corning’s ultrathin Willow glass as a substrate. The bendable panels do not yet

incorporate internal light extraction technology, so the efficacy of these is limited to around 56 lm/W. The

Brite 3 panel family is projected to ship in 2018, with rigid panels available in Q3 and bendable panels in Q4.

All Brite 3 family products are offered in a neutral (4000K) or warm (3000K) white light color, have a six-

stack tandem OLED architecture, and are designed for high brightness (7000 – 8500 cd/m2) operation. The

neutral white panels have a lower efficacy (61 lm/W for rigid panels and 44 lm/W for flexible panels) though

have a greater R9 >75 [32].

LG Display has focused on production improvements rather than on furthering their panel performance. In

2017, their catalog promised warm white (3000K) 3-stack rigid panels that would deliver 90 lm/W with CRI of

93. However, availability of these high efficacy panels was reportedly quite limited. LG has completed their

new line located at Gumi in Korea, and production of a Luflex family of panels is underway. The Luflex series

comprises panels with a range of shapes (circular, square, rectangular) and sizes as large as 30 cm x 30 cm.

Their catalog offers rigid panels in a warm (3000K) or neutral (4000K) color temperature and CRI >90. As

with the Brite 3 family products, the neutral white panels are less efficacious, offering 52 lm/W (rigid) as

compared with warm white products that reach 72 lm/W (rigid). Flexible panels are offered with warm white

emission only and have power efficacy of 50 lm/W. Advertised panel lifetime is adequate, with L70 at 3,000

cd/m2 of 40,000 hours for warm white and 30,000 hours for neutral white CCT [33].

While LG Display and OLEDWorks move forward with OLED panel technology and production, Japan’s

Lumiotec has been acquired by V-Technology and has halted production of OLED panels in order to focus on performance and production improvements [34]. Lumiotec has a range of panel products with high brightness

49

(up to 4,600 cd/m2), CCTs of 3000K, 4000K, and 5000K, and efficacies of 30–45 lm/W. Sales of panels will

continue while supplies last.

OSRAM has announced that they are exiting the OLED lighting business. In 2016, OSRAM shifted focus from

general illumination applications to making OLEDs for automotive applications. Since then, they have become

a lead supplier for Audi and BMW, which have begun incorporating OLEDs in taillights and interior lighting.

While they will continue to supply OLEDs to the automotive industry through 2020, they do not plan further

R&D efforts in OLED lighting.

China’s First-O-Lite claims some of the largest panels (37cm x 47cm) with efficacy >65 lm/W on their

second-generation OLED lighting production line. These panels reportedly have a lifetime L70 of >20,000

hours at 3,000 cd/m2 [35].

As some panel makers push forward and others pull back, the overall performance of OLEDs continues to

improve. Table 6-1 provides a breakdown of OLED panel and luminaire efficiency projections. With excellent

color quality, lifetimes of up to 100,000 hours, and efficacies approaching 100 lm/W, OLEDs specifications

are more competitive. Though costs have rapidly declined, significant reductions are needed to realize their

market potential. This section first touches on panel and luminaire efficacy projects and then describes key

R&D challenges in OLED lighting that affect the cost and performance of OLEDs. These include: 1)

performance materials for stable, efficient devices; 2) light extraction; 3) advanced fabrication technology; and

4) lighting platforms. State-of-the-art methods and new technology directions will be highlighted and targets

for future performance increases are suggested.

Table 6-1. OLED Historical and Targeted Luminaire Efficiency

Metric 2016 2018 2020 2025 Goal

Panel Efficacy1 (lm/W) 60 85 110 150 190

Optical Efficiency of Luminaire 100% 100% 100% 90%2 90%2

Efficiency of Driver 85% 85% 90% 90% 95%

Total Efficiency from Device to Luminaire 85% 85% 90% 81% 86%

Resulting Luminaire Efficacy1 (lm/W) 51 72 99 122 162

Notes:

1. Efficacy projections assume CRI >90, CCT 3000K

2. Losses representing possible use of beam shaping optics

6.1 Stable, Efficient White Organic Emitters

OLEDs have significant potential to realize energy savings through its application to large-area solid state

lighting sources. However, OLED costs and performance (e.g., device efficiency and longevity) need

improvements to stimulate adoption. While phosphorescent red and green emitter systems are available that

meet lifetime and efficiency demands, blue emitter materials and hosts have presented an ongoing challenge.

The high energy needed to create blue photons leads to the formation of excited states with energies

comparable to intramolecular bond strengths of the organic materials. These excited states decay by many non-

radiative mechanisms, leading to accelerated deterioration of the organic layers, as well as reduced efficacy.

This problem is not so severe in fluorescence since the singlet state lifetimes are short; however, in phosphorescent systems the radiative emission from triplet states is much slower, increasing the probability of

non-radiative decay and reducing device lifetime. Although many years of research at Universal Display

50

Corporation (UDC) and other companies have led to improvements in the performance of phosphorescent

emitters, current lifetime performance is not yet sufficient for commercial adoption for either lighting or

displays. To achieve practical levels of stability, commercial panels rely on fluorescent blue emitters.

Unfortunately, the efficiency of fluorescent emitters is limited to around 25% due to the ratio of singlet (25%)

to triplet (75%) states while phosphorescent emitters can achieve nearly 100% internal quantum efficiency

(IQE).

OLED display manufacturers are struggling with blue emitters and hosts as well and are investing heavily in

materials development. There are some differences in their materials requirements, such as the need for deep-

blue in displays, while sky-blue is adequate for general lighting. Additionally, general lighting applications

tend to have longer lifetime requirements than needed for displays. However, technical advancements and

materials development can likely be cross-leveraged between OLED display and general lighting. The near-

term focus (i.e., 5-10 years) for lighting is focused on small molecule materials deposited by vapor phase

deposition, which provide the best performance.12 Roll-to-roll (R2R) deposition is still of interest for OLEDs,

but vapor-phase deposition of the emissive layers can be incorporated into R2R lines. Both display and

lighting applications require low-cost materials. Though materials utilization efficiency continues to improve,

cost roadmaps demand materials prices to drop as well.

To improve the stability and efficiency of devices while also reducing costs, various alternative materials

approaches are being explored. Thermally activated delayed fluorescence (TADF) has gained the most ground

in recent years. This technology attempts to harness both singlet and triplet excitons to generate highly

efficient and stable emission of blue photons through fluorescence pathways. In molecules where the triplet

energy is close to the singlet energy, thermal upconversion of the triplet to singlet states can theoretically allow

for 100% IQE, shown in Figure 6-2. Cynora has developed sky blue (CIE = 0.37) TADF materials with an

EQE of 22% and a lifetime (L50) at 1000 (cd/m2) of greater than 1500 hours [36] [37]. They have also

reported a deep blue (CIE = 0.14) emitter with an EQE of 20% and a lifetime (L97) at 700 (cd/m2) of 20 hours.

Samsung and LG have invested $25M into Cynora for the development of alternative TADF emitters, which

are to be in mass production by 2020.

Figure 6-2. Cynora’s Illustration of TADF as Compared with Fluorescent and Phosphorescent Approaches [36]

12 There was a push for development of solution-deposited materials for cost reductions in the past few years, but such efforts are now concentrated on

display applications, which have different performance requirements and structures that are more suitable for printing.

51

There is ongoing interest in TADF materials development. In Korea, Material Science, who currently supplies

electron transport layer (ETL) and hole transport layer (HTL) materials to OLED manufacturers, has

developed blue TADF emitters and hosts for Samsung and LG Display commercial OLED production. The

National Taiwan University group has achieved record results through using an oriented TADF emitter. They

reached an EQE of ~37% for a sky-blue organic electroluminescence in a conventional planar device structure.

A highly efficient TADF emitter was used based on the spiroacridine-triazine hybrid and simultaneously

possessed nearly unitary (100%) photoluminescence quantum yield, excellent thermal stability, and strongly

horizontally-oriented emitting dipoles (with a horizontal dipole ratio of 83%) [38].

A further extension of the TADF approach has been suggested by the Kyushu University group in which two

dopants are introduced: a TADF dopant and a fluorescent dopant. Exciton formation is accomplished on the

TADF dopant, and excitons are all transferred to the singlet state of the fluorescent emitter, as displayed in

Figure 6-3. Proponents of this approach predict that device stability and efficiency can be improved over

conventional TADF because of reduced triplet energy (due to upconversion), reduced exciton lifetimes, and

more efficient transfer processes. Furthermore, this approach can take advantage of available fluorescent

emitters and is suitable for display applications as it produces the narrow spectrum of a fluorescent emitter, but

with greater efficiency. This approach has been termed “hyper-fluorescence” and is being commercialized by

Kyulux [39].

Though Kyulux’s initial objective is to develop commercial red, green, and yellow emitter/host systems, they

have reported high performance hyperfluorescent blues as well. Recent achievements include blue (470 nm

wavelength) with lifetimes (L95) at 750 cd/m2 of 100 hours and an EQE of 26 - 22% at 1000 cd/m2. Japan-

based chemical producer Nagase has invested $4.6 million in Kyulux. Kyulux’s hyperfluorescence yellow is

employed in Wisechip’s flexible passive matrix OLED (PMOLED) displays allowing the display to reduce the

power consumption by ~50% compared to a regular fluorescent PMOLED. These panels are expected to be in

mass production by the end of 2018 [40]. Hyperfluorescent technology is also being researched by a European

team (led by Merck) called the HyperOLED project, as well as by Georgia Tech in a project funded by the

BTO SSL R&D Program.

Figure 6-3. Comparison of the Mechanisms of TADF and Hyperfluorescence [41]

While they provide an alternative to phosphorescent materials, TADF approaches suffer similar lifetime

limitations due to the high energies involved and the similar order of magnitude of the excited state lifetimes.

When excited states are long-lived, there is a higher density of long-lived triplet excitons, which increases

opportunities for annihilation. In blue-emitting compounds, the energy dissipated by these exciton-quenching

reactions can be large enough to initiate molecular dissociation of the emissive material layer (EML).

Degradation of the host molecules in the emitting layer is as much of a concern as the stability of the emitter

molecule. New hosts for blue emitter systems are needed that have appropriate energy levels, charge transport

properties, and stability.

52

It is common to use emitters (phosphorescent, fluorescent, TADF) in small (<20%) doping concentrations in a

host matrix to prevent aggregation quenching. However, researchers are beginning to explore ambipolar TADF

compounds that can operate as “neat” emitter layers – composed entirely of the TADF compound [42]. Efforts

are underway at both the University of Southern California (USC) and Georgia Tech to explore this

opportunity as a portion of their BTO-funded research. Similarly, TADF molecules can be used as hosts for

phosphorescent emitters to achieve long-lived devices. In this case, the triplet excitons of the host (which are

typically unstable) are rapidly transferred to the phosphorescent dopant [43].

A hybrid fluorescent/phosphorescent approach is being explored to achieve white OLEDs at USC. In a BTO-

funded project led by Mark Thompson, the EML host transfers singlets to the blue fluorescent and triplets to

the red/green phosphorescent dopants. In this type of system, high stability and efficiency might be achieved

by harvesting the singlets and triplets independently.

In the search for new materials, the use of computational modeling has proven to be a valuable tool to help

down-select classes of materials to explore. Researchers at Merck, Harvard University in collaboration with

Kyulux, USC, and others are using such tools to predict molecular properties.

Other methods to improve materials performance and stability involve the modification of deposition

techniques. While processing conditions of organics for PV have been studied in detail, there is a paucity of

reported studies for OLED materials. Recently, TU Dresden and Universitat Autònoma de Barcelona have

presented research into the effects of deposition temperature on OLED materials properties. They found that

by evaporating the organic materials at the appropriate deposition rate and substrate temperature, they could

achieve ultra-stable glasses. Glasses evaporated at substrate temperatures around 85% of the glass transition

temperature and low growth rates (generally below 5 A/s) enhance the density – thus, the stability – of the

glass. Using four different phosphorescent emitters, over 15% increases in efficiency and operational stability

were achieved. Enhancements for TADF materials are expected to be even more impactful [44].

Panel failure can also occur due to the formation and growth of defects in the emitter layers or on the

electrodes. Experiments at Penn State University, funded by the BTO SSL Program, have shown that these

defects can arise from contaminants among the organic materials and have suggested ways to mitigate the

growth of small defects into catastrophic shorts. The BTO goal is to reduce the number of such abrupt failures

to less than 1 in 10,000.

6.2 Light Extraction

Extraction efficiency is the ratio of visible photons emitted from the panel to the photons generated in the

emissive region. For basic OLED devices on planar glass substrates, only about 20% of the generated light is

emitted from the panel. This is largely due to absorption, which is amplified by trapping of photons in the

electrodes, transparent substrates, and inner layers resulting from mismatches in the index of refraction along

the photon path from the emissive region to the outside of the device. In devices in which the cathode is

proximal to the emitting region, significant energy can also be lost through the excitation of surface plasmon

modes.

Extracting light from substrate modes can be accomplished by the use of external microlens arrays or

scattering films laminated to the transparent OLED substrate. This yields an extraction enhancement of around

1.5–1.6x, bringing the EQE of the device up to around 30 - 35%. To extract light typically lost to waveguided

modes in the anode and organic stack, internal light extraction layers can be placed between the substrate and

anode. This is a much greater challenge, considering that the additional layers threaten to complicate

manufacture and interfere with the OLED device. By incorporating both internal and external light extraction

technologies in devices, panel manufacturers can achieve as much as 2.2x extraction enhancement factors, and

EQEs >40%. Advancements in light extraction, together with refining the stack and utilizing more reflective

cathodes (where silver replaces aluminum), have led to lighting panel efficacies of 85–90 lm/W.

While this represents considerable performance enhancement as compared to previous generation devices, the

BTO target for light extraction efficiency is 75%, which corresponds to an extraction enhancement of >3.5x.

53

The extraction efficiency of current products is only 30 to 50%, leaving ample room for improvement and

energy efficiency gains. Many approaches are being explored, including: 1) scattering layers; 2) functionalized

substrates (e.g. with internal grids, lenses, gratings, corrugations) to break planar symmetry and direct light out

of the device; 3) corrugated substrates to reduce surface plasmon modes; 4) tailoring the refractive index; 5)

orientation of the emitter dipole; and 6) optimization of the OLED stack. It is important to note that there is

still significant variation in panel manufacturer’s stack structure, deposition techniques, and value proposition.

Because of this, there is likewise significant variation in the applicability of various light extraction techniques

as they depend on the OLED architecture, scale-up potential, and cost.

6.2.1 Scattering Layers

Some commercial products have incorporated scattering layers between the transparent electrode and substrate.

In films from a leading supplier, Pixelligent, nanoparticles of ZrO2 are utilized to achieve a high index polymer

matrix which also contains larger TiO2 particles as scatterers. The density of ZrO2 particles can be tailored to

achieve a graded refractive index of this layer to reduce Fresnel reflections. Using this graded index scattering

approach, extraction enhancement of up to approximately 2.5x has been reported.

Concerns with this approach are centered around the introduction of additional layers and materials to the

device. Any internal extraction film must be stable and compatible with subsequent OLED manufacturing. If

polymeric hosts are used, they must be patternable to prevent the ingress of water and oxygen through the

extraction layer to the device. Furthermore, the anode deposition and anneal temperatures can be limited and

patterning of the anode can be difficult and introduce solvents. High performance light extraction methods

(demonstrating at least 2.5 times extraction enhancement) that can be integrated into panels, without

compromising lifetime and yield, are needed.

6.2.2 Functionalized Substrates

It is difficult to increase the extraction of light from a device in which all the interfaces are planar. The

introduction of scattering particles is just one example of many strategies to add three-dimensional (3-D)

structures inside the device. Other suggestions have been to introduce grids between the emitting layers and

transparent anodes or to use internal multi-lens arrays. The latter approach was shown to be very effective in

laboratory experiments by Panasonic, but they were unable to incorporate their solution in commercial panels

[45]. Researchers at the University of Michigan explored a similar concept wherein the multi-lens array is

embedded in the substrate. With this sub-electrode microlens array approach (SEMLA), up to 70% EQE was

achieved with green OLED devices. This high efficiency was observed using an index matching fluid and large

hemispherical lens to extract as much light as possible from the substrate modes. Using the SEMLA with

external microlens arrays, EQE of around 47% for green and 27% for white OLEDs was observed, as shown in

Figure 6-4 [46].

Figure 6-4. University of Michigan Sub-electrode Microlens Array Approach: (a) Device Architecture; (b) Light Extraction

Enhancement Factor [46]

(a) (b)

54

6.2.3 Corrugated Substrates

Many researchers have suggested the use of corrugated substrates which effectively disrupts the coupling of

light to surface plasmon polariton (SPP) modes. This is being explored in multiple BTO SSL Program funded

projects. The North Carolina State University group created quasi-random grating structures with 260 nm

average period and 50 nm FWHM, where typical corrugation depth is 90nm. They observed 87% enhancement

in efficacy without any increases in leakage current [47]. The Iowa State University group have reported

enhancement factors of up to 2.4 using patterns with depth of 215 to 500 nm imprinted in polycarbonate [48].

The major problem with this approach lies in the reliability of OLEDs that are fabricated on corrugated

substrates. Corrugated substrates have been associated with electrical shorts and could contribute to local high-

field-induced degradation. It may be many years before manufacturers will accept the risk.

6.2.4 Refractive Index

The refractive index of the materials currently used in transparent substrates is close to 1.5, while that of the

emitter layer is close to 1.75. This means that much of the light does not reach the substrate and cannot be

extracted by the microlens array (MLA). The external film could be much more effective if substrates with

higher index were used. Unfortunately, no candidates have been identified on which reliable OLEDs can be

fabricated at an affordable cost. Nevertheless, the development of a set of materials with a common refractive

index would increase the effectiveness of an external MLA and would eliminate Fresnel reflections at internal

interfaces. Thus, some groups are exploring altering the index of refraction of the organic stack materials

(Professor Giebink Group at Penn State) or looking at graded index layers between the anode and substrate

(Pixelligent).

6.2.5 Orientation of Emitter Dipoles

The escape of photons is more likely when they are emitted in a direction close to the normal. This is more

likely when the molecular dipoles lie in the plane of the OLED. The development of phosphorescent layers

with oriented molecules has been pursued extensively for Ir-based emitters at the University of Southern

California and for Pt-based emitters at Arizona State University and Seoul National University [49] [50] [51].

Figure 6-5 shows that EQE over 35% can be obtained without any extraction enhancement structures [51].

Figure 6-5. External Quantum Efficiency of Phosphorescent OLEDs with Pt-Based Emitters [51]

Recently, researchers demonstrated OLEDs with as high as 56% EQE using molecular orientation and external

scattering films tailored for forward-intensive scattering [52]. By tuning characteristics of the bulk scattering

layer – such as asymmetry parameter, scattering efficiency and scatterance – the team was able to improve the effectiveness of the external scattering film. Further, their simulations show that EQEmax increases significantly

with horizontal dipole orientation, even when an external scattering layer is employed. EQEmax of an OLED

55

with perfectly oriented dipoles can reach 63%, while it is limited to 45% for isotropic orientation.

Experimental results using Ir(dmppy-ph)2tmd emitters with dipole orientation (Θ = 0.865) in combination with

SiO2 or TiO2 scattering films showed EQEs greater than 50% and as high as 56%.

In order to achieve molecular orientation of the emitter molecules, the shape of the molecule plays a large role.

Also, some studies have shown how the molecular orientation of different organic semiconductor molecules

can be tuned by changing the deposition temperature [53] [54] . In general, lower deposition temperatures lead

to more horizontal alignment.

6.2.6 Stack Optimization

Optimizing device stack structure works to minimize coupling of emission to loss modes. In addition to cavity

tuning, attention is paid to layer thicknesses and device architecture. For example, the spacing of the emissive

region from the cathode material affects the formation of surface plasmon modes. Thus, multi-stacked tandem

devices, and devices with thick ETLs will have lower losses to SPP modes. Layer thickness and materials

properties are also important to realize reduced optical absorption in OLED layers. The prevalence of multi-

stacked OLEDs and the introduction of scattering layers that recirculate many photons within the device has

led to increased concern about absorption losses. Each time that a photon is reflected back, either from the

scattering layer or the transparent substrate, it must pass across the transparent anode and organic layers twice

and then be reflected at the cathode. There are three components of special concern in this regard: the

transparent indium tin oxide (ITO) anode, charge generation layers, and the cathode.

• Transparent anode: ITO is still used in commercial OLEDs. It is extremely difficult to achieve low

sheet resistance (less than 10 Ω/sq) and low optical absorption (less than 5%) simultaneously. In the

search for alternative transparent conductors, encouraging results have been obtained in the laboratory

for silver nano-wires embedded in a polymer host, but reliable OLEDs deposited on such electrodes have

not yet been demonstrated.

• Charge generation layers: Work in a collaboration between OLEDWorks, Aixtron and RTW Aachen

University has shown that charge generation layers can lead to significant optical absorption, with

transmission rates often below 90% [55]. This loss is particularly severe in devices with six organic

stacks where it is estimated that the light extraction efficiency drops by 4% in going from 3-stack to 6-

stack structures [56].

• Cathode: Imperfect reflection at the cathode can be a major cause of photon absorption. The

OLEDWorks group demonstrated that the efficacy can be increased substantially by replacing the usual

aluminum cathode with a silver cathode. Using the silver cathode, with an internal scattering layer and

external foil, they obtained 65% light extraction in a single stack device and 57% extraction with 3

stacks. Many authors have expressed special concern about the excitation of surface plasmons in the

cathode when the emitter layer is very close to the metal electrode. The effect can be reduced by

introducing a thick electron transport layer and is of less concern in devices with multiple stacks.

6.2.7 Light Extraction Enhancement in Flexible OLEDs

Currently available internal light extraction layers are not consistent with flexible OLEDs. The major challenge

is to identify appropriate nano-particle or host materials and deposition techniques which provide layers that

are stable under bending and onto which transparent electrodes and OLEDs can be added. A second concern is

patterning of the light extraction layers to prevent the ingress of water and oxygen through the edges.

OLEDWorks and LG Display both offer flexible OLED panels with external light extraction having efficacy of

around 50 lm/W and lifetimes of 40,000 to 50,000 hours at 3000 cd/m2. OLEDWorks’ BendOLED is

fabricated on Corning’s ultrathin Willow Glass whereas LG Display switched to production on polymer

substrates due to issues with breakage of glass.

56

Figure 6-6. OLEDWorks Brite 3 Curve, BendOLED on Corning Willow Glass [57]

6.3 Advanced Fabrication Technology for OLEDs

The two major challenges with respect to the manufacture of OLED panels are to reduce cost and to enable the

production of lightweight, ultra-thin conformable panels that will lead to luminaires with distinctive form

factors.

In order to enable high-volume sales in competition with LED luminaires, the manufacturing cost of OLED

lighting panels needs to be reduced to about $100/m2. This will allow luminaires to be sold in the range of

$200/m2 to $500/m2. A path to meeting the target using traditional fabrication techniques was included in the

2017 Suggested Research Topics Supplement and is shown below in Table 6-2.

Table 6-2. Current Status and Cost Targets for Panels Produced by Traditional Methods

2016 2018 2020 2025

Substrate Area (m2) 0.2 1.2 1.2 2.7

Capital Cost ($M) 50 125 125 200

Cycle Time (minutes) 3 2 1 0.5

Capacity (1000 m2/yr) 17 175 350 2,400

Depreciation ($/m2) 600 140 70 35

Organic Materials ($/m2) 150 100 50 15

Inorganic Materials ($/m2) 200 140 100 30

Labor ($/m2) 100 25 15 5

Other Fixed Costs ($/m2) 50 15 10 5

Total (unyielded) ($/m2) 1,100 420 245 90

Yield of Good Product (%) 70 80 85 90

Total Cost ($/m2) 1,570 525 290 100

6.3.1 Depreciation Costs

Depreciation charges depend mainly on the cost of the production facility and the throughput of good products. The estimates given for 2016 in Table 6-2 were consistent with lines in production in 2016 and 2017. The only

new facility available in 2018 is the factory operated by LG Display in Gumi. With a substrate size of 1.1 m x

57

1.25 m and a design throughput of 15,000 sheets per month, the capacity of this facility could exceed that

assumed for 2018 in Table 6-2. However, production is being scaled up gradually and no commercial sales had

been announced as of mid-August 2018. The throughput achieved in the previous LG line was only 4,000

sheets per month.

It has often been suggested that roll-to-roll (R2R) processing will facilitate the reduction of cycle time.

However, cycle times of around 90 seconds are commonly achieved in the display industry by sheet-to-sheet

processing. In the short-term, the barriers to greater throughput are in the time taken for individual processes,

such as the deposition of organic materials and encapsulation.

In order to reach the long-term SSL cost targets, without the investment of the large capital sums required for

substrates as large as those now used in the display industry, it seems essential that the cycle time for the

fabrication of OLED lighting panels be reduced to well under one minute. 30 seconds remains a suitable target

for any new R&D projects.

6.3.2 Deposition of Organics

Reducing the time taken for deposition of the organic layers to less than 60 seconds is challenging. The

traditional approach to faster deposition within thermal evaporation is to increase the temperature in the

source. Stability of the organics at high temperature is one issue, as is thermal management within the

deposition chamber, especially for plastic substrates. The adoption of Organic Vapor Phase Deposition

(OVPD) or nozzle jet printing could lead to faster deposition at lower temperatures, but these techniques have

not yet been deployed commercially.

Defect control remains a substantial problem and is restricting the size of available panels. Recent evidence

from BTO-supported research at Pennsylvania State University has shown that contamination can occur during

deposition and that build-up of organic materials on the walls of the chamber must be avoided.

Data from the OLED display industry shows that the cost of organic materials may provide an obstacle to

reaching cost targets for OLED lighting. The structure of the organic layers used by LG Display in panels for

OLED TVs is similar to that used for lighting. Display Supply Chain Consultants (DSCC) has reported that the

total cost of the materials is $130/m2, with 40% attributed to the emissive layers and 60% to the transport and

charge generation layers. DSCC forecasts that the cost will decrease by only about 25% over the next four

years. Reducing the waste of organics is thus critical to achievement of the cost targets in Table 6-2.

6.3.3 Substrates and Encapsulation

DSCC estimates that the average cost of the glass used as substrates in rigid OLED displays is about $18/m2.

The cost of flexible substrates is much greater, at over $60/m2, mainly due to the need to sustain high

temperatures during formation of the active-matrix backplane. The development of less expensive substrates

for OLED lighting has achieved only limited success and further R&D is needed, especially for conformable

panels. Although toleration of very high temperatures is not required, the surface qualities of most inexpensive

plastic substrates are insufficient for OLED fabrication. The development by DuPont-Teijin of PET with a

peelable-clean surface (PCS) provides a promising new substrate that is being tested in Europe in the Lyteus

project. The addition of a SiN barrier deposited by plasma-enhanced chemical vapor deposition (PECVD) is

giving good protection against ingress of H2O and O2. The use of ultra-thin glass in conformable panels avoids

the need for a barrier, but the cost may be too high for most lighting applications.

Substantial progress has been made in techniques for encapsulation of OLED displays, using hybrid multi-

layer films. The organic layers are deposited by ink-jet printing (IJP) in a nitrogen atmosphere and the

inorganic layers by PECVD of atomic-layer deposition (ALD). However, the equipment needed for these

processes is extremely expensive, adding about $250/m2 to the depreciation cost. Less costly approaches to

encapsulation are needed for lighting applications. As a short-term solution, a thin layer of metal can be used

as a cover for bottom-emitting panels. This also helps with thermal management.

58

Edge patterning is also critical for OLED panels. The presence of layers of organic material means that H2O

and O2 can also enter through the edges of the panel, leading to early device failure. One approach is to use

printing techniques, such as IJP or slot-die coating, to ensure that no organic materials are deposited near the

edge. Alternatively, laser ablation can be used to remove the organics after deposition. Care must be taken not

to damage the fragile OLED materials that remain and to avoid contamination due to the debris.

6.3.4 Transparent Conductors

The OLED industry has long sought a replacement for ITO as a transparent conductor, partly because of its

fragility under bending and partly because of fears of price increases due to a potential shortage of indium.

Despite many years of research, a commercially viable alternative has not yet emerged. ITO can sustain the

limited bending needed for conformable lighting panels and the materials cost has not escalated. The cost and

inconvenience of acquiring patterned ITO from external suppliers has been of concern, due to the limited

availability of sources within the US. Laboratory experiments, supported by the BTO SSL program and others,

have demonstrated that alternatives such as silver nanowires and wire meshes can meet the required

performance targets, but a reliable means of fabricating patterned electrodes with a surface smooth enough for

OLED deposition has not yet been demonstrated.

One of the problems of using ITO in flexible panels is the relatively poor performance when deposited at low

temperatures. Corning has demonstrated that the use of ultra-thin glass allows deposition at up to 350o C,

achieving transparency of over 80% with sheet resistance of 12 Ω/sq.

6.3.5 Substrate Handling

Although R2R processing of OLED lighting panels has been studied in several laboratories in Asia and

Europe, there has been little activity in the US since the project at the GE Research Center in 2012. Konica

Minolta is the only company that has attempted to use R2R in high-volume production. They completed their

production facility in 2014 with a designed capacity of 1M panels per month. However, they have never

revealed the performance of the panels that can be manufactured in this plant and it is not clear that products

have been offered for commercial sales.

Although Konica Minolta has not revealed the reasons for their challenges with this approach, experience on

other pilot lines allows identification of the major challenges and assessment of progress. The most useful

information has come from the laboratories involved in the European PI-SCALE project, with pilot lines in the

Holst Center in Eindhoven and the Fraunhofer Institute in Dresden.

The team at the Holst Center has studied solution processing with both additive and subtractive patterning on

plastic and metal foils. For example, they have demonstrated that slot-die coating can be used to produce

rectangular patterns. Solvents in the deposited layers are removed by thermal curing, requiring long residence

in large ovens. They obtain good thickness control, to within +5 nm, at web speeds up to 30 m/min. The Holst

team has shown that effective hybrid barriers can be formed on PET or PEN substrates at web speeds of

4m/min, combining an organic planarization layer with a hard coat of SiNx deposited by plasma enhanced

chemical vapor deposition (PECVD).

The Fraunhofer group uses 14 vacuum deposition tools arranged around a large drum. The drum provides

cooling as well as mechanical support for the web, which can be plastic, metal or ultra-thin glass. The system

was designed to operate at web speeds of up to 1 m/min, but it was initially limited to 0.2 m/min to prevent

overheating of the substrate during metal deposition. To prevent scratching and particulate deposition during

web handling, the team strongly recommends that direct contact is limited to one side and that an interleaf

layer is inserted if partially processed webs are rolled-up before panel separation. After the OLED layers are

deposited, they can be encapsulated by lamination to a pre-prepared barrier film and are then singulated by

laser cutting. By this means, small OLEDs have been produced with yields above 80% and efficacies of 20-25

lm/W.

59

Maintaining quality control during each process step and scaling the web width from the current 300 mm to

between 1 and 1.5 m are so challenging that only relatively modest web speeds seem practical. Rates of 2-6

m/min should be adequate to reach realistic throughput targets for the next decade. Several groups have

estimated that the potential savings from R2R processing are about 20-30%. If these estimates are accurate, the

major cost reductions must come from elsewhere.

6.4 OLED Lighting Platforms

6.4.1 OLED Features

The major advantage of OLED panels is the glare-free surface lighting. Early proponents envisioned OLED

wallpaper, OLED ceilings, and OLED curtains. The luminance was kept below 3000 cd/m2, so that users could

approach the light source without shielding their eyes. However, since the cost of OLED panels scales more

directly with their size rather than the light output, the cost per lumen can be reduced by increasing the

luminance of the panels. In fact, products with luminance of up to 8000 cd/m2 have been offered. Thus, one of

the challenges in luminaire design is to balance luminance considerations and operating life against

manufacturing cost.

For luminaires that are installed above head height, glare could be reduced by beam shaping. Achieving light

distributions that differ significantly from the typical Lambertian patterns is difficult while retaining the thin

panel profile, but it is possible that some of the concepts that are being explored to enhance light extraction

could be extended to focus the beam more narrowly.

Another benefit of OLEDs is the direct production of red, green and blue light meaning that, unlike with

LEDs, there is no penalty in efficacy for delivering white light at low color temperature. Although dynamic

color tuning is possible, in principle, either by horizontal or vertical separation of the RGB sources, achieving

this without significant penalty in cost or efficacy has not yet been implemented commercially.

6.4.2 Special Applications

A number of special applications favor the intrinsic features and performance provided by OLEDs. The low

weight and thin profiles of OLED panels are especially attractive in transport applications. Much attention has

been paid to automobile tail lights, with segmented panels producing red light with luminance of 1000-2000

cd/m2. Much brighter panels will be needed for adoption in brake lights (up to 20,000 cd/m2) and direction

indicators (up to 50,000 cd/m2). Additionally, beam shaping may be needed to meet regulatory requirements

[58]. The availability of conformable white panels could also have a major impact on interior lighting for

automobiles. Although some prototypes have been demonstrated, further progress is needed in cost reduction

and assuring long lifetimes.

Weight reduction and form factor are even more important in aircraft and almost all airlines have

enthusiastically adopted SSL [59]. Some airlines offer color-tunable lighting systems, which are used to create

custom lighting scenes. Airlines may further customize their interiors through feature architecture and accent

lighting. As a critical component of airline branding, cabin lighting systems are subject to intense scrutiny,

resulting in strict color and luminance requirements. Increased reliability as measured by Mean-Time Between

Failure (MTBF) and Mean-Time Between Unscheduled Removal (MTBUR) is desirable to reduce

maintenance costs. Many parts are designed for a 20-year service life.

Similar to the automotive and aviation application examples discussed above, OLED lighting products need to

continue to seek out new and pre-existing lighting applications that are favorable for the technology and offer

meaningful energy impacts. Within the framework of lighting application efficiency, described in Section 4.1,

OLEDs may offer improved optical delivery efficiency since they can be located much close to a task surfaces

without creating glare, due to their low brightness and slim form factor.

60

6.4.3 Modules and Light Engines

Manufacturers of OLED panels and drivers have designed light engines (modules) to facilitate the market

introduction of OLED luminaires. One of the early examples was the Keuka module from OLEDWorks, which

provided a slim mounting frame and connectors for their rigid panels, along with a dimming driver. The

availability of OLED modules which include power supply electronics could enable more rapid adoption of

OLED lighting technology by simplifying the integration of OLEDs in the ultimate lighting product. The

availability of easy-to-use, cost effective, and high-performance OLED modules could also unlock new ideas

for OLED lighting products.

6.4.4 Drivers and Power Supplies

Substantial progress has been made in the development of drivers specifically for OLEDs. Dimming drivers

offering protection against shorts and open circuits are available with thickness of less than 5 mm. These can

be placed between multiple panels in conformable luminaires, but further reduction in size would be valuable.

The efficiency of DC-DC conversion is over 90%, but adapting to AC power supplies is still challenging,

partly because the required drive voltage can rise significantly during the lifetime of the panel. Cost is also of

concern, especially in luminaires providing independent control of multiple panels.

Recent progress in thin-film electronics raises the possibility of introducing an ultra-thin backplane that

contains the drivers along with sensor electronics. A prototype OLED lighting system was developed in the

European IMOLA project with a thickness of 3 mm, which the researchers believe could be reduced to around

0.3 mm [60].

61

Appendices

7.1 LED Supply Chain

Understanding and managing the manufacturing supply chain is critical to the success of any manufacturing

operation. In a general sense, the LED manufacturing processes can be defined by a sequence of relatively

independent manufacturing steps. These manufacturing steps are supported by the supply of manufacturing

equipment, materials, and testing equipment. The combination of the manufacturing processes, equipment,

materials, and testing constitute the manufacturing supply chain.

The supply chain shown in Figure 7-1 represents the current situation for LED-based SSL manufacturing, but

it should be recognized that the supply chain is ever-changing and will continue to evolve and mature. For

example, a vertically integrated manufacturer might currently handle a number of these processes internally;

however, as the manufacturing industry matures, it is common for the supply chain to become more

disaggregated for optimum manufacturing efficiency. In addition, the manufacturing supply chain will be

impacted by developments in technology and product design and can also be impacted by product distribution,

including geographical or regulatory considerations.

Figure 7-1. LED-Based SSL Manufacturing Supply Chain

Note: The blue-shaded boxes and blue arrows describe the main manufacturing flow. The supporting elements of the

supply chain are broken down into manufacturing equipment, materials, and test and measurement equipment. These

supporting elements feed into the main manufacturing flow as indicated by the relevant arrows.

The manufacturing process for LED-based luminaires begins with LED die manufacturing, consisting of

growth of the LED wafer by metal organic chemical vapor deposition (MOCVD), processing of the LED wafer

by mostly conventional semiconductor processes, and separation of the LED wafer into individual LED chips. The next step is typically to mount the LED die into LED packages, including the deposition of phosphor

material to convert the blue LED emission to white light. Finally, the LED packages are integrated with a

62

driver, heat sink, optical components, and mechanical elements to form the end luminaire or lamp product. The

manufacturing process is constantly evolving as individual elements are refined or removed, new elements are

developed, or new process sequences are introduced. Ultimately the optimum process flow for a particular

product will depend on a detailed system level optimization.

7.1.1 LED Package Manufacturing

Manufacturing Methods

The LED die manufacturing process comprises epitaxial growth of the active device layers on the substrate,

processing of the semiconductor wafer to define individual devices, dicing of the wafer to produce individual

die, and mounting of the resulting die in packages that provide mechanical support along with thermal and

electrical contacts.

The LED package no longer is the dominant cost element within the LED-based luminaire and represents a

smaller fraction of the cost, from approximately 18% in a replacement lamp to 7% or less in an LED indoor or

outdoor fixture. Efforts to reduce costs while continuing to improve performance will require concerted action

throughout the manufacturing supply chain. Such efforts will focus on higher quality and lower cost raw

materials, improved epitaxial growth equipment and processes, optimized wafer processing equipment, and

more efficient packaging methods, materials, and equipment.

There is a growing market demand for integrated light engines comprised of LEDs and the driver. The

different integration levels are illustrated in Figure 7-2. Level 1 (L1) refers to the packaged LED; Level 2 (L2)

refers to components such as LEDs or driver electronics mounted on a board; and Level 2+ (L2+) refers to

various higher levels of integration such as LEDs with optical elements. L2 and L2+ integration is desirable for

some luminaire manufacturers as it simplifies the value chain and their manufacturing process. Careful system

optimization at L2 enables the ability to tailor the LED operating conditions, optimize the number of packages

employed, and simplify the L2 configuration for lower manufacturing cost while retaining quality and

reliability. This translates to reduced system size and/or cost, which is valued by customers.

Figure 7-2. Integration Path for LED Components [61] [62] [63] [64] [65]

Package Diversity

The variety of LED packages for general illumination has exploded in recent years from a few types of

1 W class packages to a huge number of form factors, lumen levels, voltages, optical patterns, and physical

dimensions. An LED manufacturer can have as many as 50 different package families, and within each family

63

there are multiple variants based on lumen output, forward voltage, CCT, CRI, and binning tolerance. This

package diversity has given luminaire manufacturers the freedom and flexibility to use LEDs best suited for

the targeted lighting application and market.

Four main LED package platforms (shown in Figure 7-3) have emerged:

• High-power packages (1 to 5 W) typically used in products requiring small optical source size (e.g.,

directional lamps) or high reliability (e.g. street lights).

• Mid-power packages (0.1 to 0.5 W) typically used in products requiring omnidirectional emission (e.g.,

troffers, A-type lamps).

• Chip-on-board (COB) packages typically used in products needing high lumens from small optical

source or extremely high lumen density (e.g., high-bay lighting).

• Chip scale packages (CSPs), also called package-free LEDs or white chips, have gained attention as a

compact, low-cost alternative to the high-power and mid-power platforms.

Figure 7-3. Examples of High-Power, Mid-Power, Chip-on-Board, and Chip Scale LED Packages

High-power packages provide high efficacy, high luminous flux, and good reliability based on their thermal

management and optical design. The design typically consists of a large one mm2 die, or even multiple die for

a high-power array, mounted onto a ceramic substrate for thermal management. The phosphor is applied to the

chip and then a hemispherical silicone lens is over-molded onto the package. In addition to the large die, some

high-power package designs use numerous small die in series to create a high voltage package architecture

that, when grouped with a boost driver topology, can yield system efficiency improvements.

Mid-power packages were originally used in display and backlighting applications but found their way into

general lighting applications in 2012 as chip performance improvements led to viable lumen levels for general

illumination applications. Mid-power LEDs are low-cost, plastic-molded lead frame packages that typically

contain one-to-three small LED die. The die is mounted on a silver (Ag)-coated metal lead frame surrounded

by a plastic cavity. The cavity is filled with phosphor mixed in silicone to act as the down-converter and

encapsulant. Mid-power LEDs have gained favor over high-power LEDs in a number of applications due to

their low cost, which improves the lm/$ for the system.

COB arrays typically use a large array of small die mounted onto a metal-core printed circuit board (MC-PCB)

or a ceramic substrate. The LEDs are then covered with a phosphor-mixed silicone. COB arrays provide high

lumen output (up to 14,000 lumen) from a small optical source area and are typically used in high-bay lighting

and low-bay lighting. With a good thermal substrate, these COB arrays can have the same color and lumen

stability associated with high-power packages as long as the operating temperature is kept within specification.

Their ease of use in luminaire manufacturing appeals to a number of smaller luminaire manufacturers who do

not have the surface mounting equipment to assemble discrete packages onto MC-PCBs.

Chip on Board: metal-core PCB, ceramic PCB

Mid-power: leadframe, polymer package

High-power: ceramic substrates, molded lens

Chip Scale Package: flip-chip with phosphor

64

CSP LEDs have gained prominence recently due to their lower cost from minimizing materials and

manufacturing steps, as well as their small footprint allowing for tighter packing in a luminaire. The number of

CSP product offerings continues to grow, as well as the number of manufacturers offering this LED product

type. The majority of current CSP products use flip-chip die as a base, onto which the phosphor and

encapsulant is applied. Eliminating wire bonding and removing the need for lead frames or ceramic substrates

allows for a more compact size and reduced cost. The CSP manufacturers apply a conformal phosphor coating

directly onto a blue flip-chip LED die, either coated on all five sides of the chip or just on the top surface with

the side walls containing a white reflective coating.

7.1.2 LED Luminaire Manufacturing

Manufacturing Methods

Manufacturing of an LED luminaire involves combining the LEDs with mechanical and thermal components

(e.g., the heat sink), optical components to tailor the light distribution, and LED driver electronics. LED die or

packages are a critical component of all LED-based luminaires, and luminaire manufacturing revolves around

integrating the LED source with the other luminaire components to achieve the required form factor and the

optimum balance between cost, performance, product consistency, and reliability. The balance of these

features and necessary tradeoffs, depends on the lighting application, the customer profile, the incumbent

lighting performance and cost. For example, a 6” downlight for the residential market can provide 67 lm/W,

whereas a higher-end commercial downlight from the same manufacturer can reach 100 lm/W at the same

color temperature and CRI. The difference in these two models is a factor of design choices for the product

requirements for those applications. A lower-cost downlight will have fewer LEDs, which in turn are driven at

higher currents to achieve the lumen output required, thus pushing the efficacy lower due to current density

droop at higher drive currents.

Reducing the number of LEDs can lower costs at the expense of efficacy but there are further consequences to

consider: higher drive currents lead to higher temperatures in the package, which leads to earlier lumen

degradation and color shift thus affecting the luminaires reliability performance and warranty life. This

involves just one tradeoff with the LED source design. Further subsystem design choices such as heatsink,

driver, and optics designs lead to additional tradeoffs. Understanding all the nuanced performance tradeoffs

and impacts on product design and manufacturing costs, determines the efficacy, CCT, CRI, warranty life and

cost point that different luminaire products bring to market.

The fact that some form factors have lower efficacy than others does not necessarily indicate that certain LED

lighting product classes cannot be made as efficient or reliable as other LED lighting products, but instead

could reflect a specific tradeoff the manufacturer selected for the end-use case. There are certain cases, such as

etendue limited lighting designs required for narrow spot lights, that can have efficacy limitations compared to

large area light sources such as troffers (due to the small source size required to achieve small spot sizes) but it

is not fundamental in many designs.

LED-based replacement lamps and LED luminaires have a similar level of integration, but lamps use a

standard electrical interface for use within conventional lighting fixtures. Manufacturing of LED-based

lighting products shares little in common with conventional lighting products, since conventional lighting

technologies tend to be based around the fixture-plus-lamp paradigm, with the manufacturing of each part

handled completely separately, and often by separate companies. The integrated nature of an LED-based

lighting product, where fixture, light engine, and driver electronics are typically combined in a single unit,

significantly complicates the manufacturing process. Luminaire manufacturers have successfully addressed the

challenge by introducing manufacturing technologies more commonly seen in the consumer electronics

industry, simplifying the materials and manufacturing processes, introducing system-level design optimization

methodologies (including Design For Manufacturing and Design For Assembly), and developing improved

testing capabilities.

65

7.1.3 OLED Manufacturing

Although the number of companies involved in the manufacturing of OLED panels or luminaires is relatively

small, they depend on many suppliers of materials, equipment, and process techniques. The roles of the various

suppliers are indicated in Figure 7-4.

Figure 7-4. OLED-Based SSL Manufacturing Supply Chain

Figure 7-4 provides a general sense of the OLED manufacturing supply chain. However, OLED technology

and manufacturing processes need to improve in order for OLED lighting to be commercially viable in terms

of performance and cost. While the general manufacturing structure shown above is accurate, many of the

detailed processes and materials for OLED lighting are still unknown. Advancements in manufacturing

technologies cannot come at the expense of OLED device performance, which is still not quite sufficient.

Conversely, higher performance materials and device structures must be compatible with low-cost

manufacturing techniques. R&D opportunities for advanced fabrication technologies are described in more

detail in Section 6.3.

66

7.2 BTO Program Status

7.2.1 Current SSL Portfolio

The active BTO SSL R&D Portfolio13 as of March 2018, is provided in Table 7-1, including SBIR projects

which are noted with an asterisk (*). The portfolio includes 24 projects that address LED and OLED

advancements across the application spectrum. Projects balance long-term and short-term activities, as well as

large and small business, national laboratory, and university participation. The portfolio totals some $27.6

million in government and industry investment.

Figure 7-5 provides a graphical breakdown of the funding for the current SSL project portfolio as of March

2018. BTO is providing $23.0 million for the projects, and the remaining $4.6 million is cost-shared by project

awardees. Of the 26 active projects in the SSL R&D portfolio, 14 focus on LED and 10 focus on OLED

technology.

Figure 7-5. Funding of SSL R&D Project Portfolio, March 2018

BTO supports SSL R&D in partnership with industry, small business, national laboratories, and academia.

Figure 7-6 provides the approximate level of R&D funding contained in the current SSL portfolio among the

four general groups of SSL R&D partners.

13 For the full list of all current and previous DOE SSL funded projects see: https://energy.gov/eere/ssl/downloads/solid-state-lighting-project-portfolio.

https://energy.gov/eere/ssl/downloads/solid-state-lighting-project-portfolio

67

Figure 7-6. BTO SSL Total Portfolio Summary by Recipient Group, October 2018. Total funding is $23.5M.

68

Table 7-1. SSL R&D Portfolio: Current Research Projects, Q1 2018

Research Organization Project Title

LE

D

Columbia University Graded Alloy Quantum Dots for Energy Efficient Solid-State Lighting

EIE Materials (Lumenari, Inc.) Narrow Emitting Red Phosphors for Improving pcLED Efficacy

GE Global Research Highly Integrated Modular LED Luminaire

Lucent Optics* Ultra-Thin Flexible LED Lighting Panel

Lumileds Improved Radiative Recombination in AlGaInP LEDs

Lumisyn* Tunable Nanocrystal Based Phosphors with Reduced Spectral Widths

Lumisyn* Nanocrystal-based Phosphors with Enhanced Lifetime Stability

Lumisyn* New Class of Encapsulants for Blue LEDs

PhosphorTech* Hybrid Down-Converting Structures for SSL

PhosphorTech* Plasmonic Enhanced High Light Extraction Phosphor Sheets for SSL

University of California, San Diego Novel Lighting Strategies for Circadian and Sleep Health in Shift Work Applications

University of California, Santa Barbara High Performance Green LEDs for Solid-State Lighting

University of California, Santa Barbara Identification and Mitigation of Droop Mechanism in GaN-Based LEDs

Virginia Polytechnic Institute and State University

Investigating the Health Impacts of Outdoor Lighting

OLE

D

Arizona State University Stable and Efficient White OLEDS Based on a Single Emissive Material

Georgia Institute of Technology Stable White OLEDs Enabled by New Materials with Reduced Excited-State Lifetimes

Iowa State University Enhanced Light Extraction from Low Cost White OLEDs (WOLEDs) Fabricated on Novel Patterned Substrate

Luminit* Light Extraction System for OLED

North Carolina State University Low Cost Corrugated Substrates for High Efficiency OLEDs

OLEDWorks Mask-Free OLED Fabrication Process for Non-Tunable and Tunable White OLED Panels

The Pennsylvania State University Understanding, Predicting, and Mitigating Catastrophic Shorts for Improved OLED Lighting Panel Reliability

Pixelligent* Advanced Light Extraction Material for OLED Lighting

University of Michigan Eliminating Plasmon Losses in High Efficiency White OLEDs for Lighting Applications

University of Southern California Combining Fluorescence and Phosphorescence to Achieve Very Long Lifetime and Efficient White OLEDs

Cro

ss-C

utt

ing

NextEnergy Center Lighting Technology Energy Solution (LiTES)

TRC Energy Services Integrated Solutions for Optimized Performance (ISOP) Packages

Seventhwave Integrated controls package for high performance interior retrofit: field testing of packages that include lighting

Lawrence Berkeley National Laboratory

Getting Beyond Widgets: FlexLab testing of lighting technology packages


Integrated systems packages optimized for real estate life-cycle events: field testing of packages that include lighting


Integrating Autonomous HVAC and Lighting Commissioning with Fault Detection and Diagnostics (FDD) Tools

Pacific NorthWest National Laboratory Lighting field testing, modeling, validation, and stakeholder engagement

69

7.2.2 Patents

As of January 2018, 124 SSL patents have been awarded to research projects funded by DOE. Since December

2000, when DOE began funding SSL research projects, 291 patent applications have been submitted, including

those from large businesses (92), small businesses (104), universities (83), and national laboratories (12).

70

Bibliography

[1] Digital Trends, "LG’s flexible OLED lamps are ribbons of light you can twist into any shape," [Online].

Available: https://www.digitaltrends.com/home/lg-oled-lamp-news/#/1. [Accessed September 2018].

[2] Office of Management and Budget, "Memorandum for the Heads of Executive Departments and

Agencies, FY 2020 Administration Research and Development Budget Priorities," 31 July 2018.

[Online]. Available: https://www.whitehouse.gov/wp-content/uploads/2018/07/M-18-22.pdf. [Accessed

August 2018].

[3] LEDs Magazine, "Radiometric and Photometric Terms," 1 April 2005. [Online]. Available:

https://www.ledsmagazine.com/content/dam/leds/migrated/objects/features/1/1/11/CIEphotopic.jpg.

[Accessed September 2018].

[4] R. J. e. a. Lucas, "Action spectra adapted supplemental workbook," Trends Neurosci, vol. 37, no. 1,

2014.

[5] M. Roedererq, "Conjugation of monoclonal anitibodies," August 2004. [Online]. Available:

http://www.drmr.com/abcon/. [Accessed August 2018].

[6] J. Y. Tsao, M. H. Crawford, M. E. Coltrin, A. J. Fischer, D. D. Koleske, G. S. Subramania, G. T. Want,

J. J. Wierer and R. F. Karlicek, "Solid-State Lighting: Toward Smart and Ultra-efficient Solid-State

Lighting," Advanced Optical Materials, vol. 2, no. 9, pp. 809-836, 2014.

[7] E. Biery, "Creating Value Through Controls," in DOE SSL R&D Workshop, San Francisco, CA, 2015.

[8] Open Connectivity Foundation, "New Survey Highlights Importance of Interoperability in the Internet of

Things," 22 February 2017. [Online]. Available: https://openconnectivity.org/announcements/new-

survey-highlights-importance-interoperability-internet-things .

[9] American Hospital Association, "Why Interoperability Matters," [Online]. Available:

www.aha.org/content/15/interoperabilitymatters.pdf .

[10] Pacific Northwest National Laboratory, "Connected Lighting System Interoperability Study, Part 1:

Application Programming Interfaces," 2017.

[11] P. Jauregui, "Exploring and Addressing Security Risk in Smart Lighting Systems," in Strategies in Light, Las Vegas, NV, February 2015.

[12] E. Nelson, I. Wildeson and P. Deb, "Efficiency Droop in c-plane AllnGaN LEDs," in DOE SSL R&D Workshop, Raleigh, NC, 2016.

[13] Lumileds, Improved InGaN LED System Efficacy and Cost via Droop Reduction - DOE EERE Funded

Project DE-EE0007136.

[14] I. Wildeson, "Progress and outlook for III-nitride blue, green and longer wavelength direct emitters," in

DOE SSL R&D Workshop, Nashville, TN, 2018.

[15] W. Goetz, "LED Science and Technology Advancements," in DOE SSL R&D Workshop, Nashville, TN,

2018.

[16] Cree, Inc., "Cree XLamp XP-E2 Datasheet," 2018. [Online]. Available: https://www.cree.com/led-

components/media/documents/XLampXPE2.pdf.

[17] DOE SSL Program, "University of California at Santa Barbara is Determining the Origins of Efficiency

Loss in Gallium Nitride-Based LEDs Used in Lighting," 24 July 2018. [Online]. Available:

https://www.energy.gov/eere/ssl/articles/university-california-santa-barbara-determining-origins-

efficiency-loss-gallium.

[18] J. J. Wierer, N. Tansu, A. J. Fischer and J. Y. Tsao, "III-nitride quantum dots for ultra-efficient solid-

state lighting," Laser & Photonics Reviews, vol. 10, no. 4, pp. 612-622, 2016.

[19] S. Rajan, "Tunneling-based Cascaded LEDs for Efficient Solid State Lighting," in DOE SSL R&D Workshop, Raleigh, NC, 2016.

71

[20] Lumileds, "Narrow Red Phosphor Technology," 2016.

[21] J. Murphy, "Narrow-Band Emitting Phosphors for Energy Efficient SSL," in DOE SSL R&D Workshop,

Nashville, TN, 2018.

[22] Lumileds, "Lumileds 3535 HE Plus Datasheet," 2017. [Online].

[23] K. T. Shimizu, M. Bohmer, D. Estrada, S. Gangwal, S. Grabowski, H. Bechtel, E. Kang, K. J. Vampola,

D. Chamberlin, O. B. Shchekin and J. Bhardwaj, "Toward commercial realization of quantum dot based

white light-emitting diodes for general illumination," Photonics Research, vol. 5, no. 2, pp. A1-A6,

2017.

[24] LED Professional, "Quantum Dot Based White LEDs for General Illumination," 25 September 2017.

[Online]. Available: https://www.led-professional.com/resources-1/articles/quantum-dot-based-white-

leds-for-general-illumination.

[25] B. D. Mangum, T. S. Landes, B. R. Theobald and J. N. Kurtin, "Exploring the bounds of narrow-band

quantum dot downconverted LEDs," Photonics Research, vol. 5, no. 2, pp. A13-A22, 2017.

[26] P. Palomaki, "Quantum Dot Downconverters for SSL," in DOE SSL R&D Workshop, Nashville, TN,

2018.

[27] 3DPrinting.Lighting, "Philips Lighting Telecaster," 8 March 2018. [Online]. Available:

http://www.3dprinting.lighting/philips-lighting-telecaster/.

[28] LED Professional, "Additive Manufacture of Optics Goes Digital by LUXeXceL B.V.," 5 July 2016.

[Online]. Available: https://www.led-professional.com/resources-1/articles/additive-manufacture-of-

optics-goes-digital-by-luxexcel-b-v.

[29] N. Narendran, in Strategies in Light, Long Beach, CA, 2018.

[30] C. Stalker, "DOE SSL R&D Directions: Color-Tunable Lighting," in DOE SSL R&D Workshop,


[31] Pacific Northwest National Laboratory, "Gateway Demonstrations: OLED Lighting in the Offices of

Aurora Lighting Design, Inc.," March 2016. [Online]. Available:

https://energy.gov/sites/prod/files/2016/04/f30/2016_gateway_aurora-oled.pdf. [Accessed 31 May 2017].

[32] OLEDWorks, "Brite 3," [Online]. Available: https://www.oledworks.com/oled-lighting-products/brite-

3/. [Accessed September 2018].

[33] LG, "Luflex LG OLED Light," [Online]. Available: https://www.lgoledlight.com/wp-

content/uploads/down/Luflex_Eng_view.pdf. [Accessed September 2018].

[34] Lumiotec, "Information about revision of OLED panel product lineup," 23 February 2018. [Online].

Available: http://www.lumiotec.com/en/news/pdf/20180223.pdf.

[35] First O-Lite, [Online]. Available: https://fol.manufacturer.lighting/. [Accessed September 2018].

[36] L. Bergmann, "Blue high-efficiency TADF emitters for OLED lighting," in DOE SSL R&D Workshop,


[37] OLED-info, "An interview with Cynora's CMO Dr. Andreas Haldi - talking about TADF, lifetime, color

points and more," 7 August 2018. [Online]. Available: https://www.oled-info.com/interview-cynoras-

cmo-dr-andreas-haldi-talking-about-tadf-lifetime-color-points-and-more.

[38] T.-A. Lin, T. Chatterjee, W.-L. Tsai, W.-K. Lee, M.-J. Wu, M. Jiao, K.-C. Pan, C.-L. Yi, C.-L. Chung,

K.-T. Wong and C.-C. Wu, "Sky‐Blue Organic Light Emitting Diode with 37% External Quantum

Efficiency Using Thermally Activated Delayed Fluorescence from Spiroacridine‐Triazine Hybrid,"

Advanced Materials, vol. 28, no. 32, pp. 6976-6983, 2016.

[39] Kyulux, [Online]. Available: http://www.kyulux.com/technology.html. [Accessed September 2018].

[40] WiseChip Semiconductor, "Wisechip to commercialize Hyperfluorescence TADF PMOLEDs by year's

end," 2 March 2018. [Online]. Available: http://www.wisechip.com.tw/en/news-39263/Wisechip-to-

commercialize-Hyperfluorescence-TADF-PMOLEDs-by-year-s-end.html.

[41] Allled OLED, "Hyper Fluorescence," 2018. [Online]. Available: http://allled.tistory.com/50.

72

[42] X. Zhang, C. Fuentes-Hernandez, Y. Zhang, M. W. Cooper, S. Barlow, S. R. Marder and B. Kippelen,

"High performance blue-emitting organic light-emitting diodes from thermally activated delayed

fluorescence: A guest/host ratio study," Journal of Applied Physics, vol. 124, no. 5, 2018.

[43] H. Fukagawa, T. Shimizu, Y. Iwasaki and T. Yamamoto, "Operational lifetimes of organic light-emitting

diodes dominated by Förster resonance energy transfer," Scientific Reports, p. Article number: 1735,

2017.

[44] J. Rafols-Ribe, P. Will, C. Hanisch, M. Gonzalez-Silveira, S. Lenk, J. Rodriguez-Viejo and S. Reineke,

"High-performance organic light-emitting diodes comprising ultrastable glass layers," Science Advances,

vol. 4, no. 5, 2018.

[45] K. Yamae, H. Tsuji, V. Kittichungchit, Y. Matsuhisa, S. Hayashi, N. Ide and T. Komoda, "High‐

Efficiency White OLEDs with Built‐up Outcoupling Substrate," Society for Information Display,

International Symposium Digest of Technical Papers, vol. 34, no. 1, pp. 694-697, 2012.

[46] Y. Qu, J. Kim, C. Coburn and S. Forrest, "Efficient, Nonintrusive Outcoupling in Organic Light Emitting

Devices Using Embedded Microlens Arrays," ACS Photonics, vol. 5, no. 6, pp. 2453-2458, 2018.

[47] F. So, "Low cost corrugated substrates for OLEDs," in DOE SSL OLED Stakeholder Meeting, Corning,

NY, 2016.

[48] e. a. Eeshita Mann, in DOE SSL R&D Workshop, Long Beach, CA, 2017.

[49] M. Thompson, "Advances in Organic Materials for White OLEDs," in DOE SSL R&D Workshop, Long

Beach, CA, 2017.

[50] J. Li, "Efficient and Stable OLEDs for Employing Square Planar Metal Complexes," in DOE SSL OLED Stakeholder Meeting, Corning, NY, 2016.

[51] K.-H. Kim, J.-L. Liao, S. W. Lee, B. Sim, C.-K. Moon, G.-H. Lee, H. J. Kim, Y. Chi and J.-J. Kim,

"Crystal Organic Light‐Emitting Diodes with Perfectly Oriented Non‐Doped Pt‐Based Emitting Layer,"

Advanced Materials, vol. 28, no. 13, pp. 2526-2532, 2016.

[52] J. Song, K. Kim, E. Kim, C. Moon, Y. Kim, J. Kim and S. Yoo, "Lensfree OLEDs with over 50%

external quantum efficiency via external scattering and horizontally oriented emitters," Nature Communications, vol. 9, no. Article 3207, 2018.

[53] D. Yokoyama, Y. Setoguchi, A. Sakaguchi, M. Suzuki and C. Adachi, "Orientation Control of Linear‐

Shaped Molecules in Vacuum‐Deposited Organic Amorphous Films and Its Effect on Carrier

Mobilities," Advanced Functional Materials, vol. 20, no. 3, pp. 386-391, 2010.

[54] S. Dalal, D. Walters, I. Lyubimov, J. Pablo and M. Ediger, "Tunable molecular orientation and elevated

thermal stability of vapor-deposited organic semiconductors," Proceedings of the National Academy of

Sciences of the United States of America, vol. 112, no. 14, pp. 4227-4232, 2015.

[55] M. Brast, S. Axmann, M. Slawinski and M. Weingarten, "Efficient Stacked OLED processed by Organic

Vapor Phase Deposition (OVPD)," MRS Online Proceedings Library, vol. 1788, pp. 13-18, 2015.

[56] M. Boesing, "Recent Advances in OLED Lighting," in International Display Workshop, Fukuoka, Japan,

December 2016.

[57] OLEDWorks, "OLEDWorks Listed as one of EdisonReport's Top 10 Must See List," 18 May 2018.

[Online]. Available: https://www.oledworks.com/news/oledworks-listed-as-one-of-edisonreports-top-10-

must-see-list/.

[58] Audi, in OLED World Summit, San Francisco, CA, 2017.

[59] Boeing, "Integration of OLED Lighting on Aircraft," in DOE SSL R&D Workshop, Nashville, TN, 2018.

[60] A. Monte, J. Kundrata, J. Schram, M. Cauwe, A. Baric and J. Doutreloigne, "A Modular and Interactive

OLED-Based Lighting System," The Journal of the Illuminating Engineering Society, vol. 13, no. 4, pp.

211-221, 2017.

[61] Lumileds, "LUXEON T Datasheet," 2015. [Online]. Available: http://www.lumileds.com/uploads/382/DS106-pdf. [Accessed 31 May 2017].

73

[62] Lumileds, "LUXEON XR-TX Datasheet," 2016. [Online]. Available:

http://www.lumileds.com/uploads/599/DS160-pdf. [Accessed 31 May 2017].

[63] Seoul Semiconductor, "Acrich2.5 Product Data Sheet," 2 October 2015. [Online]. Available:

http://www.seoulsemicon.com/_upload/Goods_Spec/SMJQ13XXFNSX_Rev1.5.pdf. [Accessed 31 May

2017].

[64] Seoul Semiconductor, "Acrich3 Product Data Sheet," 2016. [Online]. Available:

http://www.seoulsemicon.com/_upload/Goods_Spec/%5BSPEC%5DSMJQ-XCA5W4PX_Rev1.0.pdf.

[Accessed 31 May 2017].

[65] Frost & Sullivan, "Frost & Sullivan Honours Cambridge Nanotherm for its Chip on Heat Sink

Technology for Thermal Management," 19 December 2013. [Online]. Available:

https://www.gilcommunity.com/gil-talks/bp-2013-cambridge-nanotherm-european/. [Accessed 31 May

2017].

[66] Pacific Northwest National Laboratory, "OLED Lighting Products: Capabilities, Challenges, Potential,"

May 2016. [Online]. Available: https://energy.gov/sites/prod/files/2016/08/f33/ssl_oled-

products_2016.pdf. [Accessed 30 May 2017].

74

For more information, visit:

energy.gov/eere/ssl

DOE/EE-1907 January 2019

https://www.energy.gov/eere/ssl/solid-state-lighting

Date post:	02-Feb-2020
Category:	Documents
Upload:	others
View:	1 times
Download:	0 times

2018 Solid-State Lighting R&D Opportunities · Solid-state lighting (SSL), particularly light...

Documents